Search

In this paper we consider applying economic model predictive control (EMPC) for economic optimization of a paper machine. EMPC is used to optimize overall process targets, e.g., the economy, directly in the control layer. The basic idea in EMPC is that by combining a dynamic process-model with an economic model, it is possible to predict and optimize the future economic outcome with respect to the manipulated process variables. Periodically solving such an optimization problem with updated information from measurements corresponds to a feedback controller. The results presented here are based on simulations, using a grey-box model with parameters estimated from real data, that reveal that EMPC may improve several aspects of the economic performance of a paper machine. First, EMPC may automatically prioritize among an excessive number of inputs to determine which combinations of inputs to use in order to counter disturbances in the most economically efficient manner. Also, since EMPC makes use of dynamic optimization, it may utilize control inputs with zero steady-state gain which are not used for traditional set-point tracking. Second, since EMPC is predictive in nature, it may plan ahead and prepare the process for known changes such as grade-changes, hence reducing the transition-time with a significant reduction in production loss, and thereby significant improvements in profitability, especially for machines where grade-changes are frequent. Finally, we note that EMPC typically operates the process with constraints active, as is typical for economic optimization problems in general. This may cause problems with robustness since even small exogenous disturbances or unmodelled dynamics may cause constraint violations. We therefore suggest using an adaptive approach where a constraint margin is introduced in the EMPC optimization problem to ensure that the operating point is backed off from the actual constraints relevant for production, thereby improving the robus, QC 20190418

We present ongoing work in evaluating the performance of PID-controllers under DoS attacks. The experiments are conducted in a recently developed virtual testbed, which is openly available. An important observation is that also benign physical processes may exhibit potentially dangerous oscillations under DoS attacks unless care is taken in the control implementation. An event-based PID-controller with adaptive gain shows promising performance under DoS attack., QC 20200328.QC 20210408.

Wireless sensors and networks are used only occasionally in current control loops in the process industry. With rapid developments in embedded and highperformance computing, wireless communication, and cloud technology, drastic changes in the architecture and operation of industrial automation systems seem more likely than ever. These changes are driven by ever-growing demands on production quality and flexibility. However, as discussed in "Summary," there are several research obstacles to overcome. The radio communication environment in the process industry is often troublesome, as the environment is frequently cluttered with large metal objects, moving machines and vehicles, and processes emitting radio disturbances [1], [2]. The successful deployment of a wireless control system in such an environment requires careful design of communication links and network protocols as well as robust and reconfigurable control algorithms., QC 20191108

Process optimization is used to improve the utility and the economic performance of industrial processes, and is as such central in most automation strategies. In this thesis, two feedback-based methods for online process optimization are considered: Extremum seeking control (ESC), a classic model-free method used for steady-state optimization which dates back to the early 1900's, and economic model predictive control (EMPC), a more recent method which utilizes a model to dynamically optimize the closed-loop process economics in real time. Part I of the thesis concerns ESC. Due to a well known result by Krsti\'c and Wang, it is known that the classic ESC-loop will possess a stable stationary solution in a neighborhood of the optimum when applied to dynamic plants. However, existence and stability of an optimal solution alone are not sufficient to guarantee that the ESC-loop will converge to the optimum; uniqueness also has to be considered. In this thesis, it is shown that the near-optimal solution is not necessarily unique, not even in cases where the objective, i.e., the steady-state input-output map, is convex. The stationary solutions to the loop are shown to be characterized by a condition on the local plant phase-lag, and for a biochemical reactor it is found that this condition can be satisfied not only locally at the optimum but also at arbitrary points away from the optimum. Bifurcation theory is used to show that the observed solution multiplicity may be explained by existence of fold bifurcation points, and conditions for existence of such points are given. The phase-lag condition for stationarity combined with the result by Krsti\'c and Wang suggest that the process phase-lag is connected to steady-state optimality. In this thesis, it is shown that the steady-state optimum corresponds to a bifurcation of the plant zero dynamics which is reflected in large local phase-lag variations. This explains why the classical ESC method will have a near-optimal stat, QC 20180829

Process optimization is used to improve the utility and the economic performance of industrial processes, and is as such central in most automation strategies. In this thesis, two feedback-based methods for online process optimization are considered: Extremum seeking control (ESC), a classic model-free method used for steady-state optimization which dates back to the early 1900's, and economic model predictive control (EMPC), a more recent method which utilizes a model to dynamically optimize the closed-loop process economics in real time. Part I of the thesis concerns ESC. Due to a well known result by Krsti\'c and Wang, it is known that the classic ESC-loop will possess a stable stationary solution in a neighborhood of the optimum when applied to dynamic plants. However, existence and stability of an optimal solution alone are not sufficient to guarantee that the ESC-loop will converge to the optimum; uniqueness also has to be considered. In this thesis, it is shown that the near-optimal solution is not necessarily unique, not even in cases where the objective, i.e., the steady-state input-output map, is convex. The stationary solutions to the loop are shown to be characterized by a condition on the local plant phase-lag, and for a biochemical reactor it is found that this condition can be satisfied not only locally at the optimum but also at arbitrary points away from the optimum. Bifurcation theory is used to show that the observed solution multiplicity may be explained by existence of fold bifurcation points, and conditions for existence of such points are given. The phase-lag condition for stationarity combined with the result by Krsti\'c and Wang suggest that the process phase-lag is connected to steady-state optimality. In this thesis, it is shown that the steady-state optimum corresponds to a bifurcation of the plant zero dynamics which is reflected in large local phase-lag variations. This explains why the classical ESC method will have a near-optimal stat, QC 20180829

Process optimization is used to improve the utility and the economic performance of industrial processes, and is as such central in most automation strategies. In this thesis, two feedback-based methods for online process optimization are considered: Extremum seeking control (ESC), a classic model-free method used for steady-state optimization which dates back to the early 1900's, and economic model predictive control (EMPC), a more recent method which utilizes a model to dynamically optimize the closed-loop process economics in real time. Part I of the thesis concerns ESC. Due to a well known result by Krsti\'c and Wang, it is known that the classic ESC-loop will possess a stable stationary solution in a neighborhood of the optimum when applied to dynamic plants. However, existence and stability of an optimal solution alone are not sufficient to guarantee that the ESC-loop will converge to the optimum; uniqueness also has to be considered. In this thesis, it is shown that the near-optimal solution is not necessarily unique, not even in cases where the objective, i.e., the steady-state input-output map, is convex. The stationary solutions to the loop are shown to be characterized by a condition on the local plant phase-lag, and for a biochemical reactor it is found that this condition can be satisfied not only locally at the optimum but also at arbitrary points away from the optimum. Bifurcation theory is used to show that the observed solution multiplicity may be explained by existence of fold bifurcation points, and conditions for existence of such points are given. The phase-lag condition for stationarity combined with the result by Krsti\'c and Wang suggest that the process phase-lag is connected to steady-state optimality. In this thesis, it is shown that the steady-state optimum corresponds to a bifurcation of the plant zero dynamics which is reflected in large local phase-lag variations. This explains why the classical ESC method will have a near-optimal stat, QC 20180829

Process optimization is used to improve the utility and the economic performance of industrial processes, and is as such central in most automation strategies. In this thesis, two feedback-based methods for online process optimization are considered: Extremum seeking control (ESC), a classic model-free method used for steady-state optimization which dates back to the early 1900's, and economic model predictive control (EMPC), a more recent method which utilizes a model to dynamically optimize the closed-loop process economics in real time. Part I of the thesis concerns ESC. Due to a well known result by Krsti\'c and Wang, it is known that the classic ESC-loop will possess a stable stationary solution in a neighborhood of the optimum when applied to dynamic plants. However, existence and stability of an optimal solution alone are not sufficient to guarantee that the ESC-loop will converge to the optimum; uniqueness also has to be considered. In this thesis, it is shown that the near-optimal solution is not necessarily unique, not even in cases where the objective, i.e., the steady-state input-output map, is convex. The stationary solutions to the loop are shown to be characterized by a condition on the local plant phase-lag, and for a biochemical reactor it is found that this condition can be satisfied not only locally at the optimum but also at arbitrary points away from the optimum. Bifurcation theory is used to show that the observed solution multiplicity may be explained by existence of fold bifurcation points, and conditions for existence of such points are given. The phase-lag condition for stationarity combined with the result by Krsti\'c and Wang suggest that the process phase-lag is connected to steady-state optimality. In this thesis, it is shown that the steady-state optimum corresponds to a bifurcation of the plant zero dynamics which is reflected in large local phase-lag variations. This explains why the classical ESC method will have a near-optimal stat, QC 20180829

Process optimization is used to improve the utility and the economic performance of industrial processes, and is as such central in most automation strategies. In this thesis, two feedback-based methods for online process optimization are considered: Extremum seeking control (ESC), a classic model-free method used for steady-state optimization which dates back to the early 1900's, and economic model predictive control (EMPC), a more recent method which utilizes a model to dynamically optimize the closed-loop process economics in real time. Part I of the thesis concerns ESC. Due to a well known result by Krsti\'c and Wang, it is known that the classic ESC-loop will possess a stable stationary solution in a neighborhood of the optimum when applied to dynamic plants. However, existence and stability of an optimal solution alone are not sufficient to guarantee that the ESC-loop will converge to the optimum; uniqueness also has to be considered. In this thesis, it is shown that the near-optimal solution is not necessarily unique, not even in cases where the objective, i.e., the steady-state input-output map, is convex. The stationary solutions to the loop are shown to be characterized by a condition on the local plant phase-lag, and for a biochemical reactor it is found that this condition can be satisfied not only locally at the optimum but also at arbitrary points away from the optimum. Bifurcation theory is used to show that the observed solution multiplicity may be explained by existence of fold bifurcation points, and conditions for existence of such points are given. The phase-lag condition for stationarity combined with the result by Krsti\'c and Wang suggest that the process phase-lag is connected to steady-state optimality. In this thesis, it is shown that the steady-state optimum corresponds to a bifurcation of the plant zero dynamics which is reflected in large local phase-lag variations. This explains why the classical ESC method will have a near-optimal stat, QC 20180829

In this paper, we present a hardware and software testbed designed for evaluating vehicle-to-everything (V2X) usecases. From platooning to remote driving, there are many proposals to use V2X communication to solve sustainability or safety issues in transport networks. However, researchers mostly evaluate their proposals in only simulation studies, since setting up real, full-scale field tests can often be prohibitively expensive or time-consuming. The open-sourced Small Vehicles for Autonomy (SVEA) testbed is built around a communication software stack and a 1/10th-scale automated vehicle platform suitable for both cost-effective and time-efficient experimentation with V2X use-cases. The testbed is designed to support evaluation in a wide range of conditions, such as heterogeneous networks or vehicle fleets. To illustrate the suitability of the SVEA testbed for studying V2X use-cases, we detail and implement three use-cases: platooning, adaptive speed regulation from a road-side infrastructure camera, and remote-driving by a human operator sitting in a control tower. Finally, we conclude the paper with a discussion on the use of the platform so far and future development plans., QC 20230503

This paper considers the heterogeneous traffic intersection where both Human Driven Vehicles (HDVs) and Connected and Automated Vehicles (CAVs) exist. In such a dynamic environment, CAVs must act in a way such that safety is guaranteed at all times, which is challenging due to the unpredictable nature of human behavior. To guarantee safety, in this paper we consider the worst-case behavior of HDVs by constructing the forward reachable set and ensuring collision avoidance against the forward reachable set within the CAV's planning horizon. To ensure safety at all times, a maximal invariant safe set is designed and used as a terminal constraint such that within this set there is always admissible control for CAVs to react against all possible future behavior of other vehicles safely. Finally, we propose to solve the intersection coordination problem within a Distributed Model Predictive Control (DMPC) framework where all pairwise safety constraints among CAVs are decoupled by prioritization. As a result, each CAVs solves a Mixed Integer Quadratic Programming (MIQP) problem considering collision avoidance with all CAVs of higher priority and with all HDVs. We give theoretical proof of the recursive feasibility of our proposed DMPC formulation and practical invariant safety guarantees. The resulting solution is evaluated in simulation and shows that our coordination framework can provide invariant safe coordination in a heterogeneous traffic intersection., QC 20230502

In bioprocesses, it is important to model the kinetics of the macroscopic rates of reactions since these are required to catch the dynamical aspects of a process. In [Wang et al. 2020], a modeling method involving Gaussian processes has been developed, using a kernel especially designed for the modeling of Monod-type kinetics (activation, inhibition, double component, neutral effect). However, as will be illustrated in this paper, when the number of training data is limited or the metabolite concentration data do not have large variations (which is generally the case for real-life data), this kernel can yield inaccurate models for the kinetics. In this paper, we develop a new kernel better tailored for the modeling of Monod-type kinetics and we show that it has good modeling performances in the case of a limited number of data. The idea is to use the particular structure of Monod-type functions in the design of the kernel, i.e., we incorporate prior knowledge in the modeling., QC 20221110

This thesis contributes to the field of stochastic online learning problems, with a collection of six papers each addressing unique aspects of online learning and inference problems under specific structures. The first four papers focus on exploration and inference problems, uncovering fundamental information-theoretic limits and efficient algorithms under various structures. The last two papers focus on maximizing rewards by efficiently leveraging these structures. The first paper addresses the complex problem of learning to cluster items based on binary user feedback for multiple questions. It establishes information-theoretical error lower bounds for both uniform and adaptive selection strategies under a fixed budget of rounds or users, and proposes an adaptive algorithm that efficiently allocates the budget.The second paper tackles the challenge of uncovering hidden communities in the Labeled Stochastic Block Model using single-shot observations of labels. It introduces a computationally efficient algorithm, Instance-Adaptive Clustering, which is the first to match instance-specific lower bounds on the expected number of misclassified items.The third paper delves into the best-arm identification or simple regret minimization problem within a Bayesian setting. It takes into consideration a prior distribution for the bandit problem and the expectation of simple regret with respect to that distribution, defining it as Bayesian simple regret.It characterizes the rate of Bayesian simple regret assuming certain continuity conditions on the prior, revealing that the leading term of Bayesian simple regret stems from parameters where the gap between optimal and suboptimal actions is less than . The fourth paper contributes to the fixed budget best-arm identification problem for two-arm bandits with Bernoulli rewards. It demonstrates the optim, Denna avhandling bidrar till området för stokastiska online inlärningsproblem, med en samling av sex papper som var och en behandlar unika aspekter av online inlärning och inferensproblem under specifika strukturer. De första fyra pappren fokuserar på utforskning och inferensproblem, avslöjar grundläggande informationsteoretiska gränser och effektiva algoritmer under olika strukturer. De två sista pappren fokuserar på att maximera belöningar genom att effektivt utnyttja dessa strukturer. Det första pappret behandlar det komplexa problemet att lära sig att klustra objekt baserat på binär användarfeedback för flera frågor. Det fastställer informationsteoretiska fel nedre gränser för både uniform och adaptiv urvalsstrategier under en fast budget av rundor eller användare, och föreslår en adaptiv algoritm som effektivt allokerar budgeten.Det andra pappret tar sig an utmaningen att avslöja dolda samhällen i den märkta stokastiska blockmodellen med enstaka observationer av etiketter. Det introducerar en beräkningsmässigt effektiv algoritm, Instance-Adaptive Clustering, som är den första att matcha instansspecifika nedre gränser för det förväntade antalet felklassificerade objekt.Det tredje pappret gräver djupt i problemet med bästa armidentifiering eller enkel ångerminimering inom en Bayesiansk miljö. Det tar hänsyn till en fördelning för banditproblemet och förväntan om enkel ånger med avseende på den fördelningen, vilket definierar det som Bayesiansk enkel ånger. Det karakteriserar hastigheten för Bayesiansk enkel ånger under antagande av vissa kontinuitetsvillkor på det tidigare, vilket avslöjar att den ledande termen för Bayesiansk enkel ånger kommer från parametrar där gapet mellan optimala och suboptimala handlingar är mindre än . Det fjärde pappret bidrar till det fasta budget bästa arm identifieringsproblemet för två-arm banditer med Bern, QC 20231025

Joint communication and sensing (JCAS) systems use the same spectrum, hardware and antenna resources to jointly provide spectrally efficient communication, localization and sensing services. While previous work has analyzed the performance of communication with connected objects and localization of unconnected (passive) objects, the joint positioning of both connected and passive objects is less studied. In this paper, we consider a JCAS cellular system using orthogonal frequency-division multiplexing, in which the uplink communication signal is scattered on a moving target towards multiple receiving base stations. In this setting, multistatic sensing by cooperating base stations makes it possible to position the moving target while also positioning the transmitting user equipment based on the received communication signal at the base stations. We propose a channel model that can characterize the propagation of both the communication and sensing signals, and algorithms that facilitate the estimation of direction of arrivals and range, which in turn enables the system to infer the positions of both the communicating user and the passive target. We also show some illustrative results from the algorithms that indicate what such joint positioning practically can look like., Part of ISBN 9798350311020QC 20230831

Window-opening and window-closing behaviors play an important role in indoor environmental conditions and therefore have an impact on building energy efficiency. On the other hand, the same environmental conditions drive occupants to interact with windows. Understanding this mutual relationship of interaction between occupants and the residential building is thus crucial to improve energy efficiency without disregarding occupants' comfort. This paper investigates the influence of physical environmental variables (i.e., indoor and outside climate parameters) and categorical variables (i.e., time of the day) on occupants' behavior patterns related to window operation, utilizing a multivariate logistic regression analysis. The data considered in this study are collected during winter months, when the effect on the energy consumption of the window operation is the highest, at a Swedish residential building, the KTH Live-In Lab, accommodating four occupants in separate studio apartments. Although all the occupants seem to share a sensitivity to some common factors, such as air quality and time of the day, we can also observe individual variability with respect to the most significant drivers influencing window operation behaviors., QC 20230913, Cost-and Energy-Efficient Control Systems for Buildings, CLAS–Cybersäkra lärande reglersystem, Swedish Research Council Distinguished Professor Grant, HiSS - Humanizing the Sustainable Smart City, Knut and Alice Wallenberg Foundation Wallenberg Scholar Grant

This paper considers the problem of optimizing robot navigation with respect to a time-varying objective encoded into a navigation density function. We are interested in designing state feedback control laws that lead to an almost everywhere stabilization of the closed-loop system to an equilibrium point while navigating a region optimally and safely (that is, the transient leading to the final equilibrium point is optimal and satisfies safety constraints). Though this problem has been studied in literature within many different communities, it still remains a challenging non-convex control problem. In our approach, under certain assumptions on the time-varying navigation density, we use Koopman and Perron-Frobenius Operator theoretic tools to transform the problem into a convex one in infinite dimensional decision variables. In particular, the cost function and the safety constraints in the transformed formulation become linear in these functional variables. Finally, we present some numerical examples to illustrate our approach, as well as discuss the current limitations and future extensions of our framework to accommodate a wider range of robotics applications., Part of ISBN 978-3-907144-08-4QC 20230911

This paper proposes a high-level path planning strategy under Time Petri net (TPN) formalism for a multiagent system, which is subject to Metric Interval Temporal Logic (MITL) specifications. The work aims to design a scalable model with respect to the number of agents, as the MITL formula requires multiple agents to ensure similar tasks. The obtained model is denoted Composed Time Petri net and it couples two TPN representations assigned to the motion of the agents, respectively to the MITL specification. The planning approach is based on model-checking methods and the results are evaluated on a case study applied in robotics industry., Part of ISBN 9783907144084QC 20230831

This paper considers the problem of optimizing robot navigation with respect to a time-varying objective encoded into a navigation density function. We are interested in designing state feedback control laws that lead to an almost everywhere stabilization of the closed-loop system to an equilibrium point while navigating a region optimally and safely (that is, the transient leading to the final equilibrium point is optimal and satisfies safety constraints). Though this problem has been studied in literature within many different communities, it still remains a challenging non-convex control problem. In our approach, under certain assumptions on the time-varying navigation density, we use Koopman and Perron-Frobenius Operator theoretic tools to transform the problem into a convex one in infinite dimensional decision variables. In particular, the cost function and the safety constraints in the transformed formulation become linear in these functional variables. Finally, we present some numerical examples to illustrate our approach, as well as discuss the current limitations and future extensions of our framework to accommodate a wider range of robotics applications., QC 20230831, EU H2020 CANOPIES

This paper proposes a high-level path planning strategy under Time Petri net (TPN) formalism for a multi-agent system, which is subject to Metric Interval Temporal Logic (MITL) specifications. The work aims to design a scalable model with respect to the number of agents, as the MITL formula requires multiple agents to ensure similar tasks. The obtained model is denoted Composed Time Petri net and it couples two TPN representations assigned to the motion of the agents, respectively to the MITL specification. The planning approach is based on model-checking methods and the results are evaluated on a case study applied in robotics industry., Part of ISBN 978-3-907144-08-4QC 20230911

Hydrobatic autonomous underwater vehicles (AUVs) can be efficient in range and speed, as well as agile in maneuvering. They can be beneficial in scenarios such as obstacle avoidance, inspections, docking, and under-ice operations. However, such AUVs are underactuated systems—this means exploiting the system dynamics is key to achieving elegant hydrobatic maneuvers with minimum controls. This paper explores the use of model predictive control (MPC) techniques to control underactuated AUVs in hydrobatic maneuvers and presents new simulation and experimental results with the small and hydrobatic SAM AUV. Simulations are performed using nonlinear model predictive control (NMPC) on the full AUV system to provide optimal control policies for several hydrobatic maneuvers in Matlab/Simulink. For implementation on AUV hardware in robot operating system, a linear time varying MPC (LTV-MPC) is derived from the nonlinear model to enable real-time control. In simulations, NMPC and LTV-MPC shows promising results to offer much more efficient control strategies than what can be obtained with PID and linear quadratic regulator based controllers in terms of rise-time, overshoot, steady-state error, and robustness. The LTV-MPC shows satisfactory real-time performance in experimental validation. The paper further also demonstrates experimentally that LTV-MPC can be run real-time on the AUV in performing hydrobatic maneouvers., QC 20230915

This paper proposes a method to navigate a mobile robot by estimating its state over a number of distributed sensor networks (DSNs) such that it can successively accomplish a sequence of tasks, i.e., its state enters each targeted set and stays inside no less than the desired time, under a resource-aware, time-efficient, and computation- and communication-constrained setting. We propose a new robot state estimation and navigation architecture, which integrates an event-triggered task-switching feedback controller for the robot and a two-time-scale distributed state estimator for each sensor. With the controller, the robot is able to accomplish a task by following a reference trajectory and switch to the next task when an event-triggered condition is fulfilled. With the estimator, each active sensor is able to estimate the robot state. We provide conditions to ensure that the state estimation error and the trajectory tracking deviation are upper bounded by two time-varying sequences, respectively. Furthermore, we find a sufficient condition for accomplishing a task and provide an upper bound of running time for the task. Numerical simulations of an indoor robot's localization and navigation are provided to validate the proposed architecture., Part of ISBN 978-3-907144-08-4QC 20230911

Joint communication and sensing (JCAS) is a promising technology not only to reuse spectrum and hardware resources, but also to enhance existing communication and sensing services. Recognizing these benefits of JCAS systems, several previous papers have proposed transceiver architectures, beamforming algorithms and joint target search and channel estimation schemes that facilitate the integration of communication and sensing. Realizing that localization of passive mobile objects is vital in, for example, intelligent transportation systems, this work focuses on JCAS beamforming schemes for localization of such objects. Specifically, we propose a single beam and a multibeam scheme for localization of passive objects in the coverage area. We study the inherent trade-off between sensing and communication when using these two schemes, related to the number of time slots and the transmit energy and their impact on the angle estimation errors. The results show that, in general, spatial separation between the communication and sensing beams in the multibeam scheme gives better performance compared to the temporal separation between the sensing and communication beams in the single beam scheme, especially when the power allocated to sensing is low compared to the communication power., Part of ISBN 9783800760510QC 20230831

In this paper, we consider a planning problem for a hierarchical finite state machine (HFSM) and develop an algorithm for efficiently computing optimal plans between any two states. The algorithm consists of an offline and an online step. In the offline step, one computes exit costs for each machine in the HFSM. It needs to be done only once for a given HFSM, and it is shown to have time complexity scaling linearly with the number of machines in the HFSM. In the online step, one computes an optimal plan from an initial state to a goal state, by first reducing the HFSM (using the exit costs), computing an optimal trajectory for the reduced HFSM, and then expand this trajectory to an optimal plan for the original HFSM. The time complexity is near-linearly with the depth of the HFSM. It is argued that HFSMs arise naturally for large-scale control systems, exemplified by an application where a robot moves between houses to complete tasks. We compare our algorithm with Dijkstra's algorithm on HFSMs consisting of up to 2 million states, where our algorithm outperforms the latter, being several orders of magnitude faster., Part of ISBN 978-3-907144-08-4QC 20230911

This paper presents control strategies based on time-varying convergent higher order control barrier functions for the coordination of networks of platoons. This network could be modelled by a class of leader-follower multi-agent systems, where the leaders have knowledge on the associated tasks and control the performance of their platoon involved vehicles. The followers are not aware of the tasks, and do not have any control authority to reach them. They follow their platoon leader commands for the task satisfaction. Signal temporal logic (STL) tasks are defined for the platoons coordination. Robust solutions for the task satisfaction, based on the leader's accessibility to the follower vehicles' states are suggested. In addition, using the notion of higher order barrier functions, decentralized barrier certificates for each vehicle evolving in a formation dynamic structure are proposed. Our approach finds solutions to guarantee the satisfaction of STL tasks independent of the agents' initial conditions., Part of ISBN 9798350315431QC 20230831

This paper proposes a method to navigate a mobile robot by estimating its state over a number of distributed sensor networks (DSNs) such that it can successively accomplish a sequence of tasks, i.e., its state enters each targeted set and stays inside no less than the desired time, under a resource-aware, time-efficient, and computation-and communication-constrained setting. We propose a new robot state estimation and navigation architecture, which integrates an event-triggered task-switching feedback controller for the robot and a two-time-scale distributed state estimator for each sensor. With the controller, the robot is able to accomplish a task by following a reference trajectory and switch to the next task when an event-triggered condition is fulfilled. With the estimator, each active sensor is able to estimate the robot state. We provide conditions to ensure that the state estimation error and the trajectory tracking deviation are upper bounded by two time-varying sequences, respectively. Furthermore, we find a sufficient condition for accomplishing a task and provide an upper bound of running time for the task. Numerical simulations of an indoor robot's localization and navigation are provided to validate the proposed architecture., Part of ISBN 9783907144084QC 20230831

This paper proposes Mutation-Driven Multiplicative Weights Update (M2WU) for learning an equilibrium in two-player zero-sum normal-form games and proves that it exhibits the last-iterate convergence property in both full and noisy feedback settings. In the former, players observe their exact gradient vectors of the utility functions. In the latter, they only observe the noisy gradient vectors. Even the celebrated Multiplicative Weights Update (MWU) and Optimistic MWU (OMWU) algorithms may not converge to a Nash equilibrium with noisy feedback. On the contrary, M2WU exhibits the last-iterate convergence to a stationary point near a Nash equilibrium in both feedback settings. We then prove that it converges to an exact Nash equilibrium by iteratively adapting the mutation term. We empirically confirm that M2WU outperforms MWU and OMWU in exploitability and convergence rates., QC 20230815

This paper presents control strategies based on time-varying convergent higher order control barrier functions for the coordination of networks of platoons. This network could be modelled by a class of leader-follower multi-agent systems, where the leaders have knowledge on the associated tasks and control the performance of their platoon involved vehicles. The followers are not aware of the tasks, and do not have any control authority to reach them. They follow their platoon leader commands for the task satisfaction. Signal temporal logic (STL) tasks are defined for the platoons coordination. Robust solutions for the task satisfaction, based on the leader's accessibility to the follower vehicles' states are suggested. In addition, using the notion of higher order barrier functions, decentralized barrier certificates for each vehicle evolving in a formation dynamic structure are proposed. Our approach finds solutions to guarantee the satisfaction of STL tasks independent of the agents' initial conditions., QC 20230619

This paper investigates the planning and control problems for multi-robot systems under linear temporal logic (LTL) specifications. In contrast to most of existing literature, which presumes a static and known environment, our study focuses on dynamic environments that can have unknown moving obstacles like humans walking through. Depending on whether local communication is allowed between robots, we consider two different online re-planning approaches. When local communication is allowed, we propose a local trajectory generation algorithm for each robot to resolve conflicts that are detected on-line. In the other case, i.e., no communication is allowed, we develop a model predictive controller to reactively avoid potential collisions. In both cases, task satisfaction is guaranteed whenever it is feasible. In addition, we consider the human-in-the-loop scenario where humans may additionally take control of one or multiple robots. We design a mixed initiative controller for each robot to prevent unsafe human behaviors while guarantee the LTL satisfaction. Using our previous developed ROS software package, several experiments are conductedto demonstrate the effectiveness and the applicability of the proposed strategies., Accepted paperQC 20230516

In this thesis, we investigate the design and statistical efficiency of learning algorithms in systems with a linear structure. This study is carried along three main domains, namely identification, control, and reinforcement learning, and is presented as a collection of five papers. In the first paper, we consider the problem of best-arm identification in linear bandits. We devise a simple learning algorithm based on the track-and-stop design whose sample complexity matches known instance-optimal lower bounds asymptotically. Actually, these instance-optimal lower bounds are obtained by solving an optimization problem, which in turn inspires the design of our algorithm. We also show that our algorithm is computationally competitive and performs remarkably well experimentally. In the second paper, we investigate the linear system identification problem at a finite sample size level. More precisely, we study the problem in the so-called Probably Approximately Correct (PAC) framework. We establish tight instance-specific sample complexity lower bounds via change-of-measure arguments for generic systems. In the case of stable linear dynamical systems, we derive the first non-asymptotic instance-dependent lower bounds and prove that the Least-Squares Estimator (LSE) achieves these limits in the high accuracy regime. Our analysis of the LSE is sharper, simpler, and easier to interpret than existing analyses, and relies on novel concentration results for covariates matrices with dependent data. In the third paper, we consider the problem of learning linear quadratic regulators. We devise online learning algorithms and provide guarantees on their expected regret. We consider two distinct setups, namely when the dynamics of the system are partially known or not known at all. We achieve minimal regret scalings in both setups. The algorithm we propose is a simple variant of certainty equivalence controllers, where the estimates of the dynamics are continuously updated, and, QC 20230525

This paper proposes a controller for safe lane change manoeuvres of autonomous vehicles using high-order control barrier and Lyapunov functions. The inputs are calculated using a quadratic program (CLF-CBF-QP) which admits short calculation times. The controller allows for adaptive cruise control, lane following, lane switching and ensures collision avoidance at all times. The novelty of the controller is the decentralized approach to the coordination of vehicles without switching of controllers. In particular, vehicles indicate their manoeuvres which influences their own safe region and that of neighboring vehicles. This is achieved by introducing so-called coordination functions in the design of control barrier functions. In a relevant simulation example, the controller is validated and its effectiveness is demonstrated., QC 20230511

This paper considers the problem of optimizing robot navigation with respect to a time-varying objective encoded into a navigation density function. We are interested in designing state feedback control laws that lead to an almost everywhere stabilization of the closed-loop system to an equilibrium point while navigating a region optimally and safely (that is, the transient leading to the final equilibrium point is optimal and satisfies safety constraints). Though this problem has been studied in literature within many different communities, it still remains a challenging non-convex control problem. In our approach, under certain assumptions on the time-varying navigation density, we use Koopman and Perron-Frobenius Operator theoretic tools to transform the problem into a convex one in infinite dimensional decision variables. In particular, the cost function and the safety constraints in the transformed formulation become linear in these functional variables. Finally, we present some numerical examples to illustrate our approach, as well as discuss the current limitations and future extensions of our framework to accommodate a wider range of robotics applications., QC 20230831, EU H2020 CANOPIES

This paper presents control strategies based on time-varying convergent higher order control barrier functions for the coordination of networks of platoons. This network could be modelled by a class of leader-follower multi-agent systems, where the leaders have knowledge on the associated tasks and control the performance of their platoon involved vehicles. The followers are not aware of the tasks, and do not have any control authority to reach them. They follow their platoon leader commands for the task satisfaction. Signal temporal logic (STL) tasks are defined for the platoons coordination. Robust solutions for the task satisfaction, based on the leader's accessibility to the follower vehicles' states are suggested. In addition, using the notion of higher order barrier functions, decentralized barrier certificates for each vehicle evolving in a formation dynamic structure are proposed. Our approach finds solutions to guarantee the satisfaction of STL tasks independent of the agents' initial conditions., QC 20230619

This paper investigates the planning and control problems for multi-robot systems under linear temporal logic (LTL) specifications. In contrast to most of existing literature, which presumes a static and known environment, our study focuses on dynamic environments that can have unknown moving obstacles like humans walking through. Depending on whether local communication is allowed between robots, we consider two different online re-planning approaches. When local communication is allowed, we propose a local trajectory generation algorithm for each robot to resolve conflicts that are detected on-line. In the other case, i.e., no communication is allowed, we develop a model predictive controller to reactively avoid potential collisions. In both cases, task satisfaction is guaranteed whenever it is feasible. In addition, we consider the human-in-the-loop scenario where humans may additionally take control of one or multiple robots. We design a mixed initiative controller for each robot to prevent unsafe human behaviors while guarantee the LTL satisfaction. Using our previous developed ROS software package, several experiments are conductedto demonstrate the effectiveness and the applicability of the proposed strategies., Accepted paperQC 20230516

This paper proposes a controller for safe lane change manoeuvres of autonomous vehicles using high-order control barrier and Lyapunov functions. The inputs are calculated using a quadratic program (CLF-CBF-QP) which admits short calculation times. The controller allows for adaptive cruise control, lane following, lane switching and ensures collision avoidance at all times. The novelty of the controller is the decentralized approach to the coordination of vehicles without switching of controllers. In particular, vehicles indicate their manoeuvres which influences their own safe region and that of neighboring vehicles. This is achieved by introducing so-called coordination functions in the design of control barrier functions. In a relevant simulation example, the controller is validated and its effectiveness is demonstrated., QC 20230511

This paper presents control strategies based on time-varying convergent higher order control barrier functions for the coordination of networks of platoons. This network could be modelled by a class of leader-follower multi-agent systems, where the leaders have knowledge on the associated tasks and control the performance of their platoon involved vehicles. The followers are not aware of the tasks, and do not have any control authority to reach them. They follow their platoon leader commands for the task satisfaction. Signal temporal logic (STL) tasks are defined for the platoons coordination. Robust solutions for the task satisfaction, based on the leader's accessibility to the follower vehicles' states are suggested. In addition, using the notion of higher order barrier functions, decentralized barrier certificates for each vehicle evolving in a formation dynamic structure are proposed. Our approach finds solutions to guarantee the satisfaction of STL tasks independent of the agents' initial conditions., QC 20230619

This paper investigates the planning and control problems for multi-robot systems under linear temporal logic (LTL) specifications. In contrast to most of existing literature, which presumes a static and known environment, our study focuses on dynamic environments that can have unknown moving obstacles like humans walking through. Depending on whether local communication is allowed between robots, we consider two different online re-planning approaches. When local communication is allowed, we propose a local trajectory generation algorithm for each robot to resolve conflicts that are detected on-line. In the other case, i.e., no communication is allowed, we develop a model predictive controller to reactively avoid potential collisions. In both cases, task satisfaction is guaranteed whenever it is feasible. In addition, we consider the human-in-the-loop scenario where humans may additionally take control of one or multiple robots. We design a mixed initiative controller for each robot to prevent unsafe human behaviors while guarantee the LTL satisfaction. Using our previous developed ROS software package, several experiments are conductedto demonstrate the effectiveness and the applicability of the proposed strategies., Accepted paperQC 20230516

This paper proposes a controller for safe lane change manoeuvres of autonomous vehicles using high-order control barrier and Lyapunov functions. The inputs are calculated using a quadratic program (CLF-CBF-QP) which admits short calculation times. The controller allows for adaptive cruise control, lane following, lane switching and ensures collision avoidance at all times. The novelty of the controller is the decentralized approach to the coordination of vehicles without switching of controllers. In particular, vehicles indicate their manoeuvres which influences their own safe region and that of neighboring vehicles. This is achieved by introducing so-called coordination functions in the design of control barrier functions. In a relevant simulation example, the controller is validated and its effectiveness is demonstrated., QC 20230511

This paper considers the problem of optimizing robot navigation with respect to a time-varying objective encoded into a navigation density function. We are interested in designing state feedback control laws that lead to an almost everywhere stabilization of the closed-loop system to an equilibrium point while navigating a region optimally and safely (that is, the transient leading to the final equilibrium point is optimal and satisfies safety constraints). Though this problem has been studied in literature within many different communities, it still remains a challenging non-convex control problem. In our approach, under certain assumptions on the time-varying navigation density, we use Koopman and Perron-Frobenius Operator theoretic tools to transform the problem into a convex one in infinite dimensional decision variables. In particular, the cost function and the safety constraints in the transformed formulation become linear in these functional variables. Finally, we present some numerical examples to illustrate our approach, as well as discuss the current limitations and future extensions of our framework to accommodate a wider range of robotics applications., QC 20230831, EU H2020 CANOPIES

In this thesis, we investigate the design and statistical efficiency of learning algorithms in systems with a linear structure. This study is carried along three main domains, namely identification, control, and reinforcement learning, and is presented as a collection of five papers. In the first paper, we consider the problem of best-arm identification in linear bandits. We devise a simple learning algorithm based on the track-and-stop design whose sample complexity matches known instance-optimal lower bounds asymptotically. Actually, these instance-optimal lower bounds are obtained by solving an optimization problem, which in turn inspires the design of our algorithm. We also show that our algorithm is computationally competitive and performs remarkably well experimentally. In the second paper, we investigate the linear system identification problem at a finite sample size level. More precisely, we study the problem in the so-called Probably Approximately Correct (PAC) framework. We establish tight instance-specific sample complexity lower bounds via change-of-measure arguments for generic systems. In the case of stable linear dynamical systems, we derive the first non-asymptotic instance-dependent lower bounds and prove that the Least-Squares Estimator (LSE) achieves these limits in the high accuracy regime. Our analysis of the LSE is sharper, simpler, and easier to interpret than existing analyses, and relies on novel concentration results for covariates matrices with dependent data. In the third paper, we consider the problem of learning linear quadratic regulators. We devise online learning algorithms and provide guarantees on their expected regret. We consider two distinct setups, namely when the dynamics of the system are partially known or not known at all. We achieve minimal regret scalings in both setups. The algorithm we propose is a simple variant of certainty equivalence controllers, where the estimates of the dynamics are continuously updated, and, QC 20230525

This paper presents control strategies based on time-varying convergent higher order control barrier functions for the coordination of networks of platoons. This network could be modelled by a class of leader-follower multi-agent systems, where the leaders have knowledge on the associated tasks and control the performance of their platoon involved vehicles. The followers are not aware of the tasks, and do not have any control authority to reach them. They follow their platoon leader commands for the task satisfaction. Signal temporal logic (STL) tasks are defined for the platoons coordination. Robust solutions for the task satisfaction, based on the leader's accessibility to the follower vehicles' states are suggested. In addition, using the notion of higher order barrier functions, decentralized barrier certificates for each vehicle evolving in a formation dynamic structure are proposed. Our approach finds solutions to guarantee the satisfaction of STL tasks independent of the agents' initial conditions., QC 20230619

This paper investigates the planning and control problems for multi-robot systems under linear temporal logic (LTL) specifications. In contrast to most of existing literature, which presumes a static and known environment, our study focuses on dynamic environments that can have unknown moving obstacles like humans walking through. Depending on whether local communication is allowed between robots, we consider two different online re-planning approaches. When local communication is allowed, we propose a local trajectory generation algorithm for each robot to resolve conflicts that are detected on-line. In the other case, i.e., no communication is allowed, we develop a model predictive controller to reactively avoid potential collisions. In both cases, task satisfaction is guaranteed whenever it is feasible. In addition, we consider the human-in-the-loop scenario where humans may additionally take control of one or multiple robots. We design a mixed initiative controller for each robot to prevent unsafe human behaviors while guarantee the LTL satisfaction. Using our previous developed ROS software package, several experiments are conductedto demonstrate the effectiveness and the applicability of the proposed strategies., Accepted paperQC 20230516

This paper proposes a controller for safe lane change manoeuvres of autonomous vehicles using high-order control barrier and Lyapunov functions. The inputs are calculated using a quadratic program (CLF-CBF-QP) which admits short calculation times. The controller allows for adaptive cruise control, lane following, lane switching and ensures collision avoidance at all times. The novelty of the controller is the decentralized approach to the coordination of vehicles without switching of controllers. In particular, vehicles indicate their manoeuvres which influences their own safe region and that of neighboring vehicles. This is achieved by introducing so-called coordination functions in the design of control barrier functions. In a relevant simulation example, the controller is validated and its effectiveness is demonstrated., QC 20230511

This paper considers the problem of optimizing robot navigation with respect to a time-varying objective encoded into a navigation density function. We are interested in designing state feedback control laws that lead to an almost everywhere stabilization of the closed-loop system to an equilibrium point while navigating a region optimally and safely (that is, the transient leading to the final equilibrium point is optimal and satisfies safety constraints). Though this problem has been studied in literature within many different communities, it still remains a challenging non-convex control problem. In our approach, under certain assumptions on the time-varying navigation density, we use Koopman and Perron-Frobenius Operator theoretic tools to transform the problem into a convex one in infinite dimensional decision variables. In particular, the cost function and the safety constraints in the transformed formulation become linear in these functional variables. Finally, we present some numerical examples to illustrate our approach, as well as discuss the current limitations and future extensions of our framework to accommodate a wider range of robotics applications., QC 20230831, EU H2020 CANOPIES

This paper presents control strategies based on time-varying convergent higher order control barrier functions for the coordination of networks of platoons. This network could be modelled by a class of leader-follower multi-agent systems, where the leaders have knowledge on the associated tasks and control the performance of their platoon involved vehicles. The followers are not aware of the tasks, and do not have any control authority to reach them. They follow their platoon leader commands for the task satisfaction. Signal temporal logic (STL) tasks are defined for the platoons coordination. Robust solutions for the task satisfaction, based on the leader's accessibility to the follower vehicles' states are suggested. In addition, using the notion of higher order barrier functions, decentralized barrier certificates for each vehicle evolving in a formation dynamic structure are proposed. Our approach finds solutions to guarantee the satisfaction of STL tasks independent of the agents' initial conditions., QC 20230619

This paper proposes a controller for safe lane change manoeuvres of autonomous vehicles using high-order control barrier and Lyapunov functions. The inputs are calculated using a quadratic program (CLF-CBF-QP) which admits short calculation times. The controller allows for adaptive cruise control, lane following, lane switching and ensures collision avoidance at all times. The novelty of the controller is the decentralized approach to the coordination of vehicles without switching of controllers. In particular, vehicles indicate their manoeuvres which influences their own safe region and that of neighboring vehicles. This is achieved by introducing so-called coordination functions in the design of control barrier functions. In a relevant simulation example, the controller is validated and its effectiveness is demonstrated., QC 20230511

This paper investigates the planning and control problems for multi-robot systems under linear temporal logic (LTL) specifications. In contrast to most of existing literature, which presumes a static and known environment, our study focuses on dynamic environments that can have unknown moving obstacles like humans walking through. Depending on whether local communication is allowed between robots, we consider two different online re-planning approaches. When local communication is allowed, we propose a local trajectory generation algorithm for each robot to resolve conflicts that are detected on-line. In the other case, i.e., no communication is allowed, we develop a model predictive controller to reactively avoid potential collisions. In both cases, task satisfaction is guaranteed whenever it is feasible. In addition, we consider the human-in-the-loop scenario where humans may additionally take control of one or multiple robots. We design a mixed initiative controller for each robot to prevent unsafe human behaviors while guarantee the LTL satisfaction. Using our previous developed ROS software package, several experiments are conductedto demonstrate the effectiveness and the applicability of the proposed strategies., Accepted paperQC 20230516

In this thesis, we investigate the design and statistical efficiency of learning algorithms in systems with a linear structure. This study is carried along three main domains, namely identification, control, and reinforcement learning, and is presented as a collection of five papers. In the first paper, we consider the problem of best-arm identification in linear bandits. We devise a simple learning algorithm based on the track-and-stop design whose sample complexity matches known instance-optimal lower bounds asymptotically. Actually, these instance-optimal lower bounds are obtained by solving an optimization problem, which in turn inspires the design of our algorithm. We also show that our algorithm is computationally competitive and performs remarkably well experimentally. In the second paper, we investigate the linear system identification problem at a finite sample size level. More precisely, we study the problem in the so-called Probably Approximately Correct (PAC) framework. We establish tight instance-specific sample complexity lower bounds via change-of-measure arguments for generic systems. In the case of stable linear dynamical systems, we derive the first non-asymptotic instance-dependent lower bounds and prove that the Least-Squares Estimator (LSE) achieves these limits in the high accuracy regime. Our analysis of the LSE is sharper, simpler, and easier to interpret than existing analyses, and relies on novel concentration results for covariates matrices with dependent data. In the third paper, we consider the problem of learning linear quadratic regulators. We devise online learning algorithms and provide guarantees on their expected regret. We consider two distinct setups, namely when the dynamics of the system are partially known or not known at all. We achieve minimal regret scalings in both setups. The algorithm we propose is a simple variant of certainty equivalence controllers, where the estimates of the dynamics are continuously updated, and, QC 20230525

This paper considers the problem of optimizing robot navigation with respect to a time-varying objective encoded into a navigation density function. We are interested in designing state feedback control laws that lead to an almost everywhere stabilization of the closed-loop system to an equilibrium point while navigating a region optimally and safely (that is, the transient leading to the final equilibrium point is optimal and satisfies safety constraints). Though this problem has been studied in literature within many different communities, it still remains a challenging non-convex control problem. In our approach, under certain assumptions on the time-varying navigation density, we use Koopman and Perron-Frobenius Operator theoretic tools to transform the problem into a convex one in infinite dimensional decision variables. In particular, the cost function and the safety constraints in the transformed formulation become linear in these functional variables. Finally, we present some numerical examples to illustrate our approach, as well as discuss the current limitations and future extensions of our framework to accommodate a wider range of robotics applications., QC 20230831, EU H2020 CANOPIES

In this thesis, we investigate the design and statistical efficiency of learning algorithms in systems with a linear structure. This study is carried along three main domains, namely identification, control, and reinforcement learning, and is presented as a collection of five papers. In the first paper, we consider the problem of best-arm identification in linear bandits. We devise a simple learning algorithm based on the track-and-stop design whose sample complexity matches known instance-optimal lower bounds asymptotically. Actually, these instance-optimal lower bounds are obtained by solving an optimization problem, which in turn inspires the design of our algorithm. We also show that our algorithm is computationally competitive and performs remarkably well experimentally. In the second paper, we investigate the linear system identification problem at a finite sample size level. More precisely, we study the problem in the so-called Probably Approximately Correct (PAC) framework. We establish tight instance-specific sample complexity lower bounds via change-of-measure arguments for generic systems. In the case of stable linear dynamical systems, we derive the first non-asymptotic instance-dependent lower bounds and prove that the Least-Squares Estimator (LSE) achieves these limits in the high accuracy regime. Our analysis of the LSE is sharper, simpler, and easier to interpret than existing analyses, and relies on novel concentration results for covariates matrices with dependent data. In the third paper, we consider the problem of learning linear quadratic regulators. We devise online learning algorithms and provide guarantees on their expected regret. We consider two distinct setups, namely when the dynamics of the system are partially known or not known at all. We achieve minimal regret scalings in both setups. The algorithm we propose is a simple variant of certainty equivalence controllers, where the estimates of the dynamics are continuously updated, and, QC 20230525

This paper studies the tracking control problem of networked and quantized control systems under both multiple networks and event-triggered mechanisms. Multiple networks are to connect the plant and reference system with decentralized controllers to guarantee their information transmission, whereas event-triggered mechanisms are to reduce the information transmission via multiple networks. In this paper, all networks are independent and asynchronous and have local event-triggered mechanisms, which are based on local measurements and determine whether the local measurements need to be transmitted. We first implement an emulation-based approach to develop a novel hybrid model for tracking control of networked and quantized control systems. Next, sufficient conditions are derived and decentralized event-triggered mechanisms are designed to ensure the tracking performance. Finally, a numerical example is given to illustrate the obtained results., Part of proceedings 978-94-6384-236-5QC 20220421