Search

Trajectory movement labelling is an important pre-stage for predicting connected vehicle (CV) movement at intersections. Drivers’ movement prediction and warning at intersections ensure advanced transportation safety and researchers use machine learning-based data-driven approaches to implement these technologies. However, prediction of drivers’ movements at intersections requires labelling the train and test dataset accurately with different vehicle movements at intersections to evaluate the performance of the prediction model by comparing the actual and predicted intersection movements. Moreover, due to GPS detection error or missing co-operative awareness messages (CAM), the data resides with many abnormal trajectories which are unable to be matched with regular straight or any turning movements. Especially for big data with million trajectories, it is tedious to label the movements manually. To solve this problem, we have created an automated trajectory movement classification technique using a dual approach of map matching technique and deep transfer learning modelling. Data of connected vehicle trajectory information is taken from the Ipswich Connected Vehicle Pilot (ICVP) Project, which is one of the largest connected vehicle pilots within a naturalistic driving environment in Australia. Map matching approach is performed as initial labelling by analysing the origin and destination of the vehicle CAM messages at intersections and then was converted as image datasets of 19202 samples. The map matching error and abnormal trajectories are identified by visual inspection. With properly labelled 9496 training images, 10 transfer learning models are built and tested through the remaining 9706 testing images. The maximum testing accuracy (99.73%) is achieved from the Densenet169 model, and the result shows satisfactory accuracy for individual classes: straight (99.85%), turn left (99.59), turn right (99.25), u-turn (100%), abnormal (98.63%). This model becomes a routine tool that is used daily to automatically classify thousands of trajectory movements of the C-ITS data in the ICVP project.

In this manuscript, a control strategy for electric vehicles is developed to optimise the energy consumption while respecting constraints associated with both inter-vehicle safety and comfort, which is a challenge in typical optimal control solution methodologies. Firstly, the long-term optimal control is developed using Pontryagin’s Maximum Principle (PMP). Thereafter, the obtained PMP solution is used to bound the state space for a computationally tractable Dynamic Programming (DP) optimisation to ensure the satisfaction of the safety constraints. While the acceleration subproblem solution is similar to internal combustion engine vehicles (ICEVs), the combination of regenerative and hydraulic braking significantly alters the nature of the optimal braking profile. Since a DP solution is not tractable for real-time implementation of combined braking, a fast heuristic is developed, which achieves 98% of the optimal energy recovery calculated by the DP in the simulated cases. Simulation results demonstrate that the proposed strategy respects the acceleration and safety constraints while saving approximately 5% energy use without significantly increasing travel time. Further simulations were conducted to evaluate the effect of driver preferences on energy use. It was shown that a 9.5% reduction in energy use if the driver is willing to accept a 10% speed reduction.

Automated vehicles and advanced driver-assistance systems require an accurate prediction of future traffic scene states. The tendency in recent years has been to use deep learning approaches for accurate trajectory prediction but these approaches suffer from computational complexity, dependency on a specific environment/dataset, and lack of insight into vehicle interactions. In this paper, we aim to address these limitations by proposing a Dual Learning Model (DLM) using lane occupancy and risk maps for vehicle trajectory prediction. To understand the spatial interactions of road users, make the model independent of the environment, and consider inter-vehicle distances, we embed an Occupancy Map (OM) into the trajectory prediction model. We also utilise a traffic scene Risk Map (RM) to explicitly consider a comprehensive definition of risk based on Time-to-Collision in the traffic scene. These two features employed in the encoder-decoder architecture improve system accuracy with less complexity and provide insight into the interaction between all road users. The experiment has been conducted on two different naturalistic highway driving datasets (i.e., NGSIM and HighD) demonstrating algorithm independence from a single environment. Comparison results indicate that the DLM achieves a more accurate trajectory prediction with a less complex structure compared with existing approaches in terms of RMS prediction error, which indicates the effectiveness of DLM in such a context.

Recently, emotion recognition using low-cost wearable sensors based on electroencephalogram and blood volume pulse has received much attention. Long short-term memory (LSTM) networks, a special type of recurrent neural networks, have been applied successfully to emotion classification. However, the performance of these sequence classifiers depends heavily on their hyperparameter values, and it is important to adopt an efficient method to ensure the optimal values. To address this problem, we propose a new framework to automatically optimize LSTM hyperparameters using differential evolution (DE). This is the first systematic study of hyperparameter optimization in the context of emotion classification. In this paper, we evaluate and compare the proposed framework with other state-of-the-art hyperparameter optimization methods (particle swarm optimization, simulated annealing, random search, and tree of Parzen estimators) using a new dataset collected from wearable sensors. Experimental results demonstrate that optimizing LSTM hyperparameters significantly improve the recognition rate of four-quadrant dimensional emotions with a 14% increase in accuracy. The best model based on the optimized LSTM classifier using the DE algorithm achieved 77.68% accuracy. The results also showed that evolutionary computation algorithms, particularly DE, are competitive for ensuring optimized LSTM hyperparameter values. Although DE algorithm is computationally expensive, it is less complex and offers higher diversity in finding optimal solutions.

Traffic congestion is a daily problem facing commuters in dense cities. This problem is getting worse with the rapid growth of cities’ population and migration from rural to urban areas. Recently, electric dock-less scooters have emerged as a micro mobility mode and as a potential solution for large crowded cities with limited resources. However, a question of how to charge these e-scooters has been raised. Many e-scooter companies use freelancers to charge the scooter where they compete to collect and charge the e-scooters at their homes. This competition leads the chargers to travel long distances to collect e-scooters.In this paper, we developed a mixed integer linear programming (MILP) model to solve the E-Scooter-Chargers Allocation problem. The proposed model allocates the e-scooters to the chargers with a particular emphasis on minimising the chargers’ average travelled distance to collect the e-scooters. Moreover, we modelled the charging problem as a game between two sets of disjoint players, namely e-scooters and chargers. Then we adapted the college admission algorithm (ACA) to solve the assignment problem. For the sake of comparison, we adapted the black hole optimiser (BHO) to solve this problem. The experimental results showed that ACA solutions are close to the optimal solutions found by the MILP. Furthermore, the BHO solutions are not as good as the ACA solutions. So, we recommend using the ACA to find a good solution for very large instances where MILP needs a long time to find the optimal solution.

Cooperative Intelligent Transportation Systems (C-ITS) are being deployed in several cities around the world. Evaluation of their safety benefits in Field Operational Tests (FOT) is needed to demonstrate its benefits and build public awareness and uptake. The result of the evaluation can tell us if the C-ITS algorithms that trigger the safety events, and consequently the Human-Machine Interface (HMI) messages, are appropriately fine-tuned to induce the expected driver behavior change formulated as hypothesis. In this paper we will introduce the safety hypotheses for seven driver safety use cases being deployed by the Queensland Department of Transport and Main Roads in the Ipswich Connected Vehicle Pilot in Australia. The safety performance indicators to test these hypotheses will be introduced as well. The main challenge in evaluating the safety benefits is only using the information collected from C-ITS units without augmentation from any other sensors. To validate data collection, we ran an experiment at the Mt Cotton training facility close to Brisbane and collected Cooperative Awareness Message (CAM) messages to analyze them and check whether the speed and acceleration information extracted from them is accurate enough to detect change in speed and braking. The analysis results show that we can detect the change in speed, but the acceleration/braking pattern is very noisy and needs careful manipulation to retrieve the braking behavior.

A new Ecological and Safe (EcoSafe) driving control algorithm has been recently developed by the authors for controlling the longitudinal motion of the vehicle to minimize fuel consumption while respecting safety constraints. The algorithm uses a Model predictive control framework augmented with enhanced safety constraints based on Intervehicular Time (TIV) and the Time to Collision (TTC). This algorithm requires tuning to adapt to traffic condition. In this paper we propose a tuning method for EcoSafe algorithm which is deduced from driver preference and traffic flow information. In addition, to the best of our knowledge, the benefits of similar EcoSafe algorithms have not been tested with naturalistic data. Hence, we assessed the benefits of EcoSafe algorithm in terms of eco-driving and safety by using 1,100 km of naturalistic driving data. We use velocity profile extracted from the Australian Naturalistic Driving Study (ANDS) as the leading vehicle driving behaviour. The results show that our proposed strategy has a 14% reduction in fuel consumption on average while maintaining high safety levels without increasing travel time significantly.

Cooperative Systems provide, through the multiplication of information sources over the road, a lot of potential to improve the assessment of the road risk describing a particular driving situation. In this paper, we compare the performance of a cooperative risk assessment approach against a non-cooperative approach; we used an advanced simulation framework, allowing for accurate and detailed, close-to-reality simulations. Risk is estimated, in both cases, with combinations of indicators based on the TTC. For the noncooperative approach, vehicles are equipped only with an AAC-like forward-facing ranging sensor. On the other hand, for the cooperative approach, vehicles share information through 802.11p IVC and create an augmented map representing their environment; risk indicators are then extracted from this map. Our system shows that the cooperative risk assessment provides a systematic increase of forward warning to most of the vehicles involved in a freeway emergency braking scenario, compared to a non-cooperative system.

Simulation has been widely used to estimate the benefits of Cooperative Systems (CS) based on Inter-Vehicular Communications (IVC). This paper presents a new architecture built with the SiVIC simulator and the RTMaps™ multisensors prototyping platform. We introduce several improvements from a previous similar architecture, regarding IVC modelisation and vehicles' control. It has been tuned with on-road measurements to improve fidelity. We discuss the results of a freeway emergency braking scenario (EEBL) implemented to validate our architecture's capabilities.

IEEE 802.11p is the new standard for inter-vehicular communications (IVC) using the 5.9 GHz frequency band; it is planned to be widely deployed to enable cooperative systems. 802.11p uses and performance have been studied theoretically and in simulations over the past years. Unfortunately, many of these results have not been confirmed by on-tracks experimentation. In this paper, we describe field trials of 802.11p technology with our test vehicles. Metrics such as maximum range, latency and frame loss are examined.

Inter-Vehicular Communications (IVC) are considered a promising technological approach for enhancing transportation safety and improving highway efficiency. Previous theoretical work has demonstrated the benefits of IVC in vehicles strings. Simulations of partially IVC-equipped vehicles strings showed that only a small equipment ratio is sufficient to drastically reduce the number of head on collisions. However, these results are based on the assumptions that IVC exhibit lossless and instantaneous messages transmission. This paper presents the research design of an empirical measurement of a vehicles string, with the goal of highlighting the constraints introduced by the actual characteristics of communication devices. A warning message diffusion system based on IEEE 802.11 wireless technology was developed for an emergency breaking scenario. Preliminary results are presented as well, showing the latencies introduced by using 802.11a and discussing early findings and experimental limitations.

Cooperative systems, also known as V2I and V2V-based, are becoming increasingly important in meeting the demand for safer and more advanced driving assistance systems. Research on wireless networks to support inter-vehicular communications has made great advances. Cooperative systems will significantly enhance and increase driver's ability to be aware, anticipate and react to crash risk situations. Provision of such enhanced capabilities must meet the wellestablished features of existing wireless network systems, such as robust scalability and reliability but, more importantly cater for driver's needs, cognitive and physiological limitations. There has been much work in the area of driver's situation awareness in recent years. Situation awareness is the primary basis for subsequent driver's decision-making and performance in the operation of complex driving tasks such as maneuvering or anticipation. However there is still a gap between the two research fields despite their potentials to improve road safety. This talk will highlight interdisciplinary research issues at the frontier of cooperative systems and human cognition. The focus will be on drivers' situational awareness needs/capabilities and how cooperative systems can address them with the view to improve road safety.

Our social life is characterised by norms that manifest as attitudinal and behavioural uniformities among people. With greater awareness about our social context, we can interact more efficiently. Any theory or account of human interaction that fails to include social concepts could be suggested to lack a critical element. This paper identifies social concepts that need to be supported by future context-aware systems. It discusses the limitations of existing context-aware and Multi-Agent Systems (MAS) to support social psychology theories related to the identification and membership of social groups. We argue that social norms are among the core modeling concepts that future context-aware systems need to capture with the view to support and enhance social interactions. The social concepts identified in this paper could be used to simulate agent interactions imbued with social norms or use ICT to facilitate, assist or enhance social interactions. They also could be used in virtual communities modeling where the awareness of a community as well as the process of joining and exiting a community are important.

International audience; Nowadays, more and more driving assistances are available to help the driver and to improve the vehicle handling. With increasing sensing capacities, it also becomes possible to have a local view of the vehicle surrounding. Hence, the next steps are to provide the driving assistances a supervisor that schedules all the possible actions and plane trajectories. In this article, we propose a decision method that fits the driver decision and action schemes. The system is at three levels, for action, decision and long range planning. the article is focused on the second layer and the interaction with other layer. The second layer algorithm, based on risk assumption, evaluates, in the vehicle vicinity, the risk related to each detected object. It then computes possible actions for lower level layer. The developed method is constrained by the future implementation on a vehicle : low computation time available and small memory size.

Automobiles have deeply impacted the way in which people travel but they have also caused a number of deaths and injury due to crashes. Driver inattention, inexperience, poor judgment and/or fatigue etc., are the major contributing factors in crashes. In order to address these complex issues, scientists have developed a number of technological systems ranging from anti-lock braking system (ABS), early warning systems to complete autonomous driving solutions. Driverpsilas driving competencies also play a vital role in judging the road environment and reacting in-time to avoid any possible collision. Therefore driverpsilas perceptual and motor skills remain a key factor impacting on-road safety. Currently, there is a need for researchers to develop and evaluate intelligent transport systems (ITS) that assess driving competencies. These tools should be designed to help driver instructors and drivers to accurately evaluate the driving behaviour. This paper outlines a vision for future research in the area of intelligent driver training systems for enhancing road safety and also presents preliminary implementation.

It is a great pleasure to welcome you to the 2008 IPS/UMS/WNGS/MUS workshop held in conjunction with the International Conference on Multimedia and Ubiquitous Engineering (MUE2008), Hanwha Resort Haeundae, Busan, Korea. The workshop covers topics on ubiquitous multimedia services and security in ubiquitous multimedia systems. All the submitted papers have undergone rigorous review process by the technical program committee members and external reviewers. We have selected 17 best papers for presentation and publication in the conference proceedings.

The crash rate in road intersection demonstrates the need for a fast and accurate collision detection system. Ubiquitous computing research provides a significant opportunity to develop novel ways of improving road intersection safety. The existing intersection collision warning or avoidance systems are mostly built to suit a particular intersection. We suggest that an intersection collision detection system should be able to adapt to different types of intersections by acquiring the collision patterns of the intersection through data mining. Collision patterns that are specific to that intersection are stored in a knowledge base to select vehicles which are exposed to a high risk of collision. This algorithm increases the speed of collision detection calculation, as detection is not applied on all possible pairs in an intersection. The performance and accuracy of the algorithm are evaluated. This evaluation is done on a developed simulation bed and the results are presented.

An intersection safety system should adapt to the particular characteristics that identify an intersection, by mining traffic and collision data. Given the large amount of sensor data that are obtained for intersections and from sensor-equipped cars, analysis and learning of such data is essential. This paper presents a new method to improve safety at intersections using a combination of a mathematical based collision detection algorithm and data mining. A number of scenarios at a simulated intersection are explored with encouraging results from our data mining implementation. The results suggest that our approach can help improve situation awareness and automate understanding of intersections, which, in turn, can be used to increase safety at intersections.

Technology based driver training is still in its infancy. There is a strong need for improved and integrated computer-based screening tools in order to facilitate objective and reliable driver training assessments. This paper describes a system that analyses videos of driving sessions collected by on-board web-cameras. The system detects and tracks lane markings in order to estimate the relative position of the vehicle with respect to its lane. The system is computationally efficient as it exploits fully the off-line nature of the data. The analysis of the video recording is performed in reverse temporal order. The benefits of this approach compared to the forward analysis traditionally used are an improved robustness and a lower computational cost.

The ever increasing use of ICT (information communication technology) in cars to improve the mobility, comfort and safety has sparkled the interests of the ambient intelligent research community. It is widely accepted that the use of advanced in-vehicle technology could contribute to a driver's distraction. Information overload or inappropriateness of the methods and time to convey information to the driver are among the potential drawbacks of the use of in-vehicle technology. This paper identifies the future research directions on driving assistance interactions design. It focuses on factors that preserve or enhance the driver's ability to drive and presents an in-vehicle ambient intelligent transport (I-VAITS) architecture.

In this contribution we present a visual driver surveillance system to monitor the driver's head motion as well as the eye blink patterns. Based on these measured features the system is able to detect symptoms of fatigue and monotony. The main advantages of the presented system in contrast to existing ones is the usage of standard equipment to achieve a good cost-performance ratio, fast computation time, the possibility of measurements in darkness and the consideration of monotony. The image analysis is realized in a coarse-to-fine architecture. At first the driver's face is detected which is based on a boosted cascade of Haar wavelets. Then the eyes are searched in the face and occurring eye blinks measured by analyzing the optical flow of the eyes' region. The performance of the system was tested successfully under ideal and natural conditions.

The ever increasing use of ICT (Information Communication Technology) in cars to improve mobility, comfort and safety has sparkled interests of Ambient Intelligent research community. It is widely accepted that the use of advanced in-vehicle technology could contribute to driver’s distraction. Information overload or inappropriateness of the methods and time to convey information to the driver are among the potential drawbacks of the use of in-vehicle technology. This paper identifies future research directions on driving assistance interactions design. It focuses on factors that preserve or enhance driver’s ability to drive and presents an In-Vehicle Ambient Intelligent Transport (IVAITS) architecture.

The world of meat faces a permanent need for new methods of meat quality evaluation. Recent advances in the area of computer and video processing have created new ways to monitor quality in the food industry. Here, the authors describe an image processing technique to determine fat content in beef meat. To achieve this NMR (Nuclear Magnetic Resonance) images of beef meat have been used. The inherent advantages of NMR images are many. Chief among these are unprecedented contrasts between the various structures present in meat, as muscle, fat, and connective tissue. Moreover, the three-dimensional nature of the NMR method allows the analysis of isolated cross-sectional slices of the meat and the measure of volumetric content of fat, no only the superficially visible one. The authors propose a segmentation algorithm for the detection of fat and a filtering technique to remove intensity inhomogeneities in NMR images, cause by non-uniformities of magnetic field during acquisition. Measurements have been successfully correlated with chemical analysis and digital photography. The authors also propose a method to quantify the distribution of fat. Their results show that NMR technique is a promising non-invasive method to determine fat content in meat.

Networking homes, offices, cars and hand-held computers is the current trend of distributed mobile computing. The ever growing demand of the enterprise for integrating new technologies and new services requires a generic infrastructure. An infrastructure that allows its components to be accessed, packaged, self described, dynamically configurable, moved over wired or wireless network, deployed either in a workstation or a mobile device from anywhere at anytime. An infrastructure that can be used with new alternatives to current human-machine interfaces applicable in diverse environment.