Search

We describe a multi-modal software system for executing navigation missions in an urban environment, focusing on the robust treatment of anomalous situations such as blocked roads, stalled vehicles and tight maneuvering. Various recovery mechanisms are described relative to the nominal mode of operation, and results are discussed from the system’s deployment in the DARPA Urban Challenge.

The objective of the CyberScout project is to develop an autonomous surveillance and reconnaissance system. In this paper, we focus on advances in vision-based surveillance agents for detection, scene mosaicing, classification and correspondence. An agent-based software framework is used to promote synergy between the various surveillance algorithms and provide a distributed computing infrastructure for the system.

We are developing a Sensor Web-relevant system called the Telesupervised Adaptive Ocean Sensor Fleet that uses a group of National Oceanic and Atmospheric Administration extended-deployment autonomous surface vehicles to enable in-situ study of surface and sub-surface characteristics of Harmful Algal Blooms (HAB). The architecture supports adaptive reconfiguration based on environmental sensor inputs (“smart” sensing), and increases data-gathering effectiveness and science return while reducing demands on scientists for tasking, control, and monitoring. It combines and adapts prior related work done at Carnegie Mellon University, NASA Goddard Space Flight Center, Wallops Flight Facility, Emergent Space Technologies, and the Jet Propulsion Laboratory. Initial multi-vessel HAB characterization tests will be performed during summer 2007 with rhodamine dye as a HAB simulant and an airborne sensor validation system. The described architecture is broadly applicable to ecological forecasting, water management, carbon management, disaster management, coastal management, homeland security, and planetary exploration.

This paper describes the development of an autonomous robotic vehicle for tactical distributed surveillance under the CyberScout project at Carnegie Mellon University’s Institute for Complex Engineered Systems. The primary aim of the CyberScout project is to develop novel mobile and unattended ground sensor platforms that will extend the sphere of awareness and mobility of small military units. The research is focused on developing algorithms for multi-agent collaboration, efficient perception, sensor fusion, heterogeneous swarm technologies, distributed command and control, and task decomposition. The mobile robotic platforms within CyberScout range in size from microrobots (cm) to All Terrain Vehicles (ATV). In this paper, the hardware system, control architecture, sensor suite, current capabilities, future research, and applications for the robotic ATV (CyberATV) are described.

For autonomous vehicles, navigation systems must be accurate enough to provide lane-level localization. Highaccuracy sensors are available but not cost-effective for production use. Although prone to significant error in poor circumstances, even low-cost GPS systems are able to correct Inertial Navigation Systems (INS) to limit the effects of dead reckoning error over short periods between sufficiently accurate GPS updates. Kalman filters (KF) are a standard approach for GPS/INS integration, but require careful tuning in order to achieve quality results. This creates a motivation for a KF which is able to adapt to different sensors and circumstances on its own. Typically for adaptive filters, either the process (Q) or measurement (R) noise covariance matrix is adapted, and the other is fixed to values estimated a priori. We show that by adapting Q based on the state-correction sequence and R based on GPS receiver-reported standard deviation, our filter reduces GPS root-mean-squared error by 23% in comparison to raw GPS, with 15% from only adapting R.

A good simulation environment will facilitate motion planning algorithm development for urban autonomous driving. The first requirement of such a simulator is to be able to replicate a complex urban environment, including road network, curb, general objects, etc. The second requirement is to simulate a realistic host vehicle, which includes perception, control and vehicle dynamics, to recreate imperfect inputs and non-accurate execution of the planner. The third requirement is to model traffic participants (other on-road vehicles) for microscopic traffic simulation. Intelligent-agent-based techniques are used to allow the traffic participants to interact with the environment and each other. In this paper, we present an open-source lightweight simulation environment, FastSim, which is designed to meet the three requirements above.

Multi-robot systems (MRS) have great promise for revolutionizing the way a variety of important and complex tasks are performed. While the underlying science is advancing quickly, the engineering problems associated with deploying multi-robot team under real world constraints have not been adequately addressed. In this work, we are developing teams of Cooperative Robotic Watercraft (CRW) for critical applications including flood response, water monitoring and security. This paper details the steps in the development and deployment of our low-cost, robust CRW system, including design considerations, system description, user interface, and subsequent field-testing results. We took the watercraft into real environments and ran them through the types of exercises they will perform in real deployments, to better understand the full range of issues involved in creating and deploying real multi-robot systems. We report field testing results from two unique and different environments: four days of testing in an irrigation pond and six weeks in the Philippines, including after a typhoon; resulting in more than 100 boat hours in the water and hundreds of thousands of data points. By the end, the process and the resultant boats were effective and robust, and could be controlled by one non-computer science undergraduate student and local Filipinos with no formal education.

Recent research in robot exploration and mapping has focused on sampling environmental hotspot fields. This exploration task is formalized by Low, Dolan, and Khosla (2008) in a sequential decision-theoretic planning under uncertainty framework called MASP. The time complexity of solving MASP approximately depends on the map resolution, which limits its use in large-scale, high-resolution exploration and mapping. To alleviate this computational difficulty, this paper presents an information-theoretic approach to MASP (iMASP) for efficient adaptive path planning; by reformulating the cost-minimizing iMASP as a reward-maximizing problem, its time complexity becomes independent of map resolution and is less sensitive to increasing robot team size as demonstrated both theoretically and empirically. Using the reward-maximizing dual, we derive a novel adaptive variant of maximum entropy sampling, thus improving the induced exploration policy performance. It also allows us to establish theoretical bounds quantifying the performance advantage of optimal adaptive over non-adaptive policies and the performance quality of approximately optimal vs. optimal adaptive policies. We show analytically and empirically the superior performance of iMASP-based policies for sampling the log-Gaussian process to that of policies for the widely-used Gaussian process in mapping the hotspot field. Lastly, we provide sufficient conditions that, when met, guarantee adaptivity has no benefit under an assumed environment model., Comment: 19th International Conference on Automated Planning and Scheduling (ICAPS 2009), Extended version with proofs, 11 pages

A successful plan for space exploration requires the commissioning of fleets of robots to prospect, mine, build, inspect and maintain structures, and generally assist astronauts, rendering the overall mission as safe as reasonably achievable for human beings, the most precious resource. The authors are currently developing, under the support of NASA, a Robot Supervision Architecture (RSA) which will allow a small number of human operators to safely and efficiently telesupervise a fleet of autonomous robots. This represents a significant advance over the state of the art, where currently one robot is overseen by a group of skilled professionals. In this paper we describe some aspects of this work, including the architecture itself for coordination of human and robot work, failure and contingency management, high-fidelity telepresence, and operation under limited bandwidth. We also present highlights of our first application: wide area prospecting of minerals and water in support of sustained outposts on the Moon and on Mars.

This paper proposes the use of on-line scheduling in human telesupervision of robot teams to determine and maximize the number of robots a human operator can manage at once. To maximize the number of robots in a team, the double- Shifted Shortest Processing Time (dSSPT) algorithm was proposed for this purpose. This scheduling algorithm has the objective of minimizing total downtime (or flow time) for the group. Simulated experimental results show that dSSPT: 1) reduces downtime with respect to existing scheduling algorithms; and 2) increases the achievable spanof- control (i.e., the number of robots supervised) in a human-multirobot team compared to an existing method of prediction [3], especially as the variation in task lengths increases.

Prospecting for in situ mineral resources is essential for establishing settlements on the Moon and Mars. To reduce human effort and risk, it is desirable to build robotic systems to perform this prospecting. An important issue in designing such systems is the sampling strategy: how do the robots choose where to prospect next? This paper argues that a strategy called Adaptive Cluster Sampling (ACS) has a number of desirable properties: compared to conventional strategies, (1) it reduces the total mission time and energy consumption of a team of robots, and (2) returns a higher mineral yield and more information about the prospected region by directing exploration towards areas of high mineral density, thus providing detailed maps of the boundaries of such areas. Due to the adaptive nature of the sampling scheme, it is not immediately obvious how the resulting sampled data can be used to provide an unbiased, low-variance estimate of the regional mineral density. This paper therefore investigates new mineral density estimators, which have lower error than previously-developed estimators; they are derived from the older estimators via a process called Rao-Blackwellization. Since the efficiency of estimators depends on the type of mineralogical population sampled, the population characteristics that favor ACS estimators are also analyzed. The ACS scheme and our new estimators are evaluated empirically in a detailed simulation of the prospecting task, and the quantitative results show that our approach can yield more minerals with less resources and provide more accurate mineral density estimates than previous methods.

Mobile robots are notoriously unreliable. In order to make significant improvements in mobile robot reliability, we need quantitative methods and precise language for measuring and discussing reliability. Such methods exist within the reliability engineering literature but have seen little use in the design of mobile robots. In this report we present an overview of reliability engineering methods which can be used to predict the probability of failure for mobile robots. We also present here a novel extension to the concept of mean time to failure that incorporates the effects of operating conditions on failure rate. Finally, we demonstrate how these techniques can be used in mobile robot mission design to predict the probability of completing mission tasks. Using these methods, a mission designer can make informed decisions trading reliability against other variables such as cost.

This paper presents a method using stochastic simulation to evaluate the reliability of robot teams consisting of modular robots. For an example planetary exploration mission we use this method to compare the performance of a repairable robot team with spare modules versus nonrepairable robot teams. Our results show that for this mission a repairable robot team can provide a higher probability of mission completion than a nonrepairable team, even when the nonrepairable robots are built using components with an order of magnitude greater reliability than the repairable robots.

Decentralized multirobot learning refers to the use of multiple learning entities to achieve the optimal solution for the overall robot system. We demonstrate that single-robot learning theory can be successfully used with multirobot systems, but with certain conditions. The success and the effectiveness of this method are potentially affected by various factors that we classify into two groups: the nature of the robots and the nature of the learning entities. Incorrect setup of these factors may lead to undesirable results. In this paper, we methodically test the effect of varying four common factors (reward scope, learning algorithms, diversity of robots, and number of robots) in a decentralized multirobot system, first in simulation and then on real robots. The results show that two of these factors, reward scope and learning algorithm, if set up incorrectly, can prevent optimal, cooperative solutions.

In Carnegie Mellon University’s CyberScout project, we are developing mobile and stationary sentries capable of autonomous reconnaissance and surveillance. In this paper, we describe recent advances in the areas of efficient perception algorithms (detection, classification, and correspondence) and mission planning. In detection, we have achieved improved rejection of camera jitter and environmental variations (e.g., lighting, moving foliage) through multi-modal filtering, and we have implemented panoramic backgrounding through pseudo-real-time mosaicing. In classification, we present methods for discriminating between individuals, groups of individuals, and vehicles, and between individuals with and without backpacks. In correspondence, we describe an accurate multi-hypothesis approach based on both motion and appearance. Finally, in mission planning, we describe mapbuilding using multiple sensory cues and a computationally efficient decentralized planner for multiple platforms.

This paper describes an adaptive fuzzy throttle control for an All Terrain Vehicle (ATV) powered by internal combustion engine. The design objective is to provide smooth throttle movement, zero steady-state speed error, and to maintain a selected vehicle speed over varying road slopes for a 2MPH to 30MPH speed range. Unlike modern production vehicles which have microprocessor based engine management systems, the ATV’s engine is mechanically controlled via a carburetor. A complete mathematical model of the engine is not available making it very difficult to apply convention control techniques. Using experience and data collected from extensive experiments conducted on the ATV throttle mechanism, an adaptive fuzzy throttle control algorithm is designed. A candidate Lyapunov function is employed in the adaptive law synthesis to ensure convergence. Experimental results are presented showing the effectiveness of the control algorithm at speeds below 2.74MPH (1.2m/s).

This paper describes our current effort to develop robotic vehicles for tactical distributed surveillance. Our research is focused on multi-agent collaboration, reconfigurable systems, efficient perception and sensor fusion, distributed command and control, and task decomposition. In particular, this paper describes the main features and capabilities of our All Terrain Vehicles (ATVs), concentrating on their autonomous navigation capabilities.

An important class of robotic applications potentially involves multiple, cooperating robots: security or military surveillance, rescue, mining, etc. One of the main challenges in this area is effective cooperative control: how does one determine and orchestrate individual robot behaviors which result in a desired group behavior? Cognitive (planning) approaches allow for explicit coordination between robots, but suffer from high computational demands and a need for a priori, detailed world models. Purely reactive approaches such as that of Brooks are efficient, but lack a mechanism for global control and learning. Neither approach by itself provides a formalism capable of a sufficiently rapid and rich range of cooperative behaviors. Although we accept the usefulness of the reactive paradigm in building up complex behaviors from simple ones, we seek to extend and modify it in several ways. First, rather than restricting primitive behaviors to fixed input-output relationships, we include memory and learning through feedback adaptation of behaviors. Second, rather than a fixed priority of behaviors, our priorities are implicit: they vary depending on environmental stimuli. Finally, we scale this modified reactive architecture to apply not only for an individual robot, but also at the level of multiple cooperating robots: at this level, individual robots are like individual behaviors which combine to achieve a desired aggregate behavior. In this paper, we describe our proposed architecture and its current implementation. The application of particular interest to us is the control of a team of mobile robots cooperating to perform area surveillance and target acquisition and tracking.