Md Abdus Samad Kamal | Gunma University (original) (raw)
Papers by Md Abdus Samad Kamal
2007 IEEE International Conference on Control Applications, 2007
Sometimes a driver deviates from his natural or normal driving style due to inadequate attention ... more Sometimes a driver deviates from his natural or normal driving style due to inadequate attention or faces abnormal situation caused by a number of psychological and physical factors. Such abnormalities often lead a driver to a mistake that may cause an accident. This paper presents a novel approach called driver-adaptive assist system to avoid such abnormalities in driving scenario as a preventive measure against occurrence of vehicle collisions, assuming that natural driving style of individual drivers is the safest style. Adaptive fuzzy system with statistics of recent fluctuations records are used to determine the driving behavior from noisy data. Another fuzzy reasoning section determines the level of abnormality in driving to notify or warn the driver so that he can pay back his full concentration in driving. Different simulated drivers with pseudo realistic styles in starting, stopping and car following are used to investigate performance of the proposed system. Empirical results show the ability of the system to recognize abnormality of drivers having different driving styles.
2021 IEEE Intelligent Vehicles Symposium (IV), 2021
The state prediction of the preceding vehicle (PV) is essential for anticipatory driving that ena... more The state prediction of the preceding vehicle (PV) is essential for anticipatory driving that enables a vehicle to take early steps for efficient driving. Such driving results in energy efficiency and traffic flow improvement but also requires repeated rigorous computations. Here we propose a Deep Neural Network (DNN) based driving scheme (DDS) that is trained to predict the PV and decide the control input of a host vehicle (HV) using two DNN models. We use the Road-Speed Profile (RSP), which provides approximated traffic speeds along the road using information from limited connected vehicles, to predict the future speed of the PV. Next, with the predicted speed of the PV, the control input is instantly determined to drive the HV efficiently. For making the optimal control decisions, the DNN is trained using a Model Predictive Control (MPC) scheme that generates the control input of the HV by minimizing a typical objective function reflecting costs related to inefficient driving. In the simulation, it is found that the proposed DDS guides the HV to change acceleration early in the optimized way by predicting the PV, particularly in changing traffic conditions, which results in a significant reduction in fuel consumption of the HV. Most remarkably, the DDS can determine the decision almost in real-time, which removes the necessity of setting a long step in the discrete-time control framework as found in MPC to match with its computation time. The proposed DDS is evaluated in some extreme cases and compared with traditional driving.
Journal of Robotics and Mechatronics, 2014
<div class=&am... more <div class=""abs_img""><img src=""[disp_template_path]/JRM/abst-image/00260005/09.jpg"" width=""300"" />Network with four intersections</div> In this paper a network-wide traffic signal control scheme in a model predictive control framework using mixed integer programming is presented. A concise model of traffic is proposed to describe a signalized road network considering conservation of traffic. In the model, the traffic of two sections that belong to a traffic signal group of a junction are represented by a single continuous variable. Therefore, the number of variables required to describe traffic in the network becomes half compared with the models that describe section wise traffic flows. The traffic signal at the junction is represented by a binary variable to express a signal state either green or red. The proposed model is transformed into a mixed logical dynamical system to describe the traffic flows in a finite horizon, and traffic signals are optimized using mixed integer linear programming (MILP) for a given performance index. The scheme simultaneously optimizes all traffic signals in a network in the context of model predictive control by successively extending or terminating a green or red signal of each junction. Consequently, traffic signal patterns with the optimal free parameters, i.e., the cycle times, the split times and the offsets, are realized. Use of the proposed concise traffic model significantly reduces the computation time of the scheme without compromising the performance as it is evaluated on a small road network and compared with a previously proposed scheme. </span>
Autonomous robots are robots which can perform desired tasks in unstructured environments without... more Autonomous robots are robots which can perform desired tasks in unstructured environments without continuous human guidance or intervention. All kinds of robots have some degree of autonomy. Different robots can be autonomous in different ways or from different angle point of view. In other word, Autonomy is the other way of describing the level of intelligence any machine has. A fully autonomous robot should have the ability to perform some basic tasks such as gaining information about the environment, work for an ...
SAE Technical Paper Series, 2015
IEEE Transactions on Vehicular Technology, 2022
Recently developed efficient driving schemes usually solve a predictive optimization problem or d... more Recently developed efficient driving schemes usually solve a predictive optimization problem or determining the vehicle control input, and at the expense of high computational cost, they improve the overall traffic flows and individual driving performances on urban roads. This paper presents a more practical technique for automated vehicles’ predictive driving by extending the existing adaptive cruise control (ACC) scheme with a look-ahead functionality. Such a look-ahead driving scheme (LDS) predicts the states of the preceding vehicle at an adaptive look-ahead time step and, with negligible computation costs, computes the vehicle control input more circumspectly for efficient driving in urban traffic. The proposed LDS is evaluated in typical urban traffic at the signalized intersections by observing the intersection utilization, flowing characteristics, and individual vehicles’ fuel efficiency. Furthermore, we also evaluate the influences of the LDS-vehicles’ penetration rates on overall traffic performances at various traffic volumes.
Electronics
Human facial emotion recognition (FER) has attracted the attention of the research community for ... more Human facial emotion recognition (FER) has attracted the attention of the research community for its promising applications. Mapping different facial expressions to the respective emotional states are the main task in FER. The classical FER consists of two major steps: feature extraction and emotion recognition. Currently, the Deep Neural Networks, especially the Convolutional Neural Network (CNN), is widely used in FER by virtue of its inherent feature extraction mechanism from images. Several works have been reported on CNN with only a few layers to resolve FER problems. However, standard shallow CNNs with straightforward learning schemes have limited feature extraction capability to capture emotion information from high-resolution images. A notable drawback of the most existing methods is that they consider only the frontal images (i.e., ignore profile views for convenience), although the profile views taken from different angles are important for a practical FER system. For deve...
A reinforcement learning algorithm is proposed that can cope with high dimensionality for a class... more A reinforcement learning algorithm is proposed that can cope with high dimensionality for a class of problems with symmetrical actions. The action selection does not need considering all the states but only needs looking at a part of the states. Moreover, every symmetrical action is related to the same kind of part of state, and thus the value function can be shared, which greatly reduces the reinforcement learning problem size. The overall learning algorithm is equivalent to the standard reinforcement learning algorithm. Simulation results and other aspects are compared with standard and other reinforcement learning algorithms. Reduction in dimensionality, much faster convergence without worsening other objectives show the effectiveness of the proposed mechanism on a high dimensional optimization problem having symmetrical actions.
2015 IEEE Intelligent Vehicles Symposium (IV), 2015
Anticipative control of vehicles is a potential approach for improving travel efficiency of indiv... more Anticipative control of vehicles is a potential approach for improving travel efficiency of individual vehicles, smoothing traffic flows on urban roads, alleviating impacts on the environment and elevating comforts of the users in various respects. This paper presents such a vehicle driving system in a model predictive control (MPC) framework to efficiently drive a vehicle on multi-lane roads. Anticipation enhances the driving intelligence and strengthens the vehicle's ability in taking advance action, e.g., lane change, speed adjustment, in a dynamically varying traffic environment. More elaborately, presuming a connected vehicle environment, the system receives the information form the surrounding vehicles and infrastructure instantly through V2X communication systems and, using dynamical models, predicts the future road-traffic states. Considering relevant constraints and a performance index, the system generates the optimal acceleration and executes lane change maneuver opti...
ABSTRACT Multiagent reinforcement learning system requires an efficient but simple way of coopera... more ABSTRACT Multiagent reinforcement learning system requires an efficient but simple way of cooperation among the autonomous agents for their successful operation in the environment. This paper presents a new approach of indirect cooperation making the agents perform the task in a competitive manner. One part of the task is performed according to a task-oriented common Q-table, which represents the information of a passive element, how the agent should work with it. The other part of the task, only related to the agent movement is done according to the agent's individual Q-table. This method converges and found very effective for dynamic environment where the goal is changed after certain episodes. Learning can be done fast and effectively even if the initialization of agents and the passive elements are random in the environment for each episode, and failure is reduced greatly.
Ieej Transactions on Electronics, Information and Systems, 2005
In this paper, a reinforcement learning method is proposed that optimizes passenger service in el... more In this paper, a reinforcement learning method is proposed that optimizes passenger service in elevator group systems. Task-oriented reinforcement learning using multiple agents is applied in the control system in allocating immediate landing calls to the elevators and operating them intelligently in attaining better service in this stochastic dynamic domain. The proposed system shows better adaptive performance in different traffic
Abstract—Sometimes a driver deviates from his natural or normal driving style due to inadequate a... more Abstract—Sometimes a driver deviates from his natural or normal driving style due to inadequate attention or faces abnormal situation caused by a number of psychological and physical factors. Such abnormalities often lead a driver to a mistake that may cause an accident. This paper presents a novel approach called driver-adaptive assist system to avoid such ab-normalities in driving scenario as a preventive measure against occurrence of vehicle collisions, assuming that natural driving style of individual drivers is the safest style. Adaptive fuzzy system with statistics of recent fluctuations records are used to determine the driving behavior from noisy data. Another fuzzy reasoning section determines the level of abnormality in driving to notify or warn the driver so that he can pay back his full concentration in driving. Different simulated drivers with pseudo realistic styles in starting, stopping and car following are used to investigate performance of the proposed system. Empi...
Multiagent reinforcement learning system requires an efficient but simple way of cooperation amon... more Multiagent reinforcement learning system requires an efficient but simple way of cooperation among the autonomous agents for their successful operation in the environment. This paper presents a new approach of indirect cooperation making the agents perform the task in a competitive manner. One part of the task is performed according to a task-oriented common Q-table, which represents the information of a passive element, how the agent should work with it. The other part of the task, only...
IET Intelligent Transport Systems, 2010
This study presents a novel concept of an ecological driver assistance system (EDAS) that may pla... more This study presents a novel concept of an ecological driver assistance system (EDAS) that may play an important role in intelligent transportation systems (ITS) in the near future. The proposed EDAS is designed to measure relevant information of instant vehicle–road–traffic utilising advanced sensing and communication technologies. Using models of vehicle dynamics and traffic flow, it anticipates future situations of the vehicle–road–traffic network, estimates fuel consumption and generates the optimal control input necessary for ecological driving. Once the optimal control input becomes available, it could be used to assist the driver through a suitable human interface. The vehicle control method is developed using model predictive control algorithm with a suitable performance index to ensure safe and fuel-efficient driving. The performance of the EDAS, in terms of speed behaviour and fuel consumption, is evaluated on the microscopic transport simulator AIMSUN NG. Comparative results are graphically illustrated and analysed to signify the prospect of the proposed EDAS in building environmentally friendly ITS.
2020 59th Annual Conference of the Society of Instrument and Control Engineers of Japan (SICE), 2020
Energy consumption of vehicles at signalized intersections is highly influenced by acceleration/d... more Energy consumption of vehicles at signalized intersections is highly influenced by acceleration/deceleration maneuvers and idling time. Existing research on fuel-efficient driving at signalized intersections mainly focused on connected and automated vehicle (CAV) technologies. This paper introduces a sustainable eco-driving strategy learning scheme from driving data of a vehicle that generates the optimal speed trajectory when approaching a signalized intersection. The goal of the proposed system is to reduce fuel consumption by advising an appropriate driving strategy near a signalized intersection. Our main contribution is that the driving performance in traffic scenarios in the existing signalized intersection can be improved without utilization of vehicle-to-vehicle (V2V) and infrastructure-to-vehicle (I2V) communications. A Gaussian process regression (GPR) model is developed that predicts intersection crossing time and crossing probability of a vehicle from its driving data. B...
2017 IEEE 20th International Conference on Intelligent Transportation Systems (ITSC), 2017
This paper addresses a partially connected vehicle environment (PCVE), where a only fraction of v... more This paper addresses a partially connected vehicle environment (PCVE), where a only fraction of vehicles are equipped with V2V communication technology, and presents a method of realizing highly anticipative driving in it. More specifically, anticipative driving refers to the predictive control of a host vehicle considering the preceding traffic conditions in an extended view but in an abstract form. For acquiring the traffic conditions of the preceding road section, a road-speed profile is proposed, which is dynamically updated using limited V2V data in a bi-level approximation framework. In the lower level of the framework, the state of the known vehicles within the communication range are predicted in a short horizon using a conditional persistence prediction model. The model parameter is determined by fitting with the experimental driving data. In the upper level, the time, position and speed of the vehicles obtained from the lower level are then mapped onto the road, which is r...
World Congress, 2011
This paper presents an ecological (eco) driving system based on prediction of the preceding vehic... more This paper presents an ecological (eco) driving system based on prediction of the preceding vehicle using model predictive control. At any measured road-traffic states it computes the optimal vehicle control input using the models of the vehicle dynamics and fuel consumption. The prediction model the preceding vehicle is formulated based on experimentally obtained driving data. The proposed system is evaluated for driving on urban roads containing traffic control signals at the intersections using the microscopic transport simulator AIMSUN. Significant improvement in fuel efficiency by introducing the model of the preceding vehicle has been confirmed from the simulation results.
2007 IEEE International Conference on Control Applications, 2007
Sometimes a driver deviates from his natural or normal driving style due to inadequate attention ... more Sometimes a driver deviates from his natural or normal driving style due to inadequate attention or faces abnormal situation caused by a number of psychological and physical factors. Such abnormalities often lead a driver to a mistake that may cause an accident. This paper presents a novel approach called driver-adaptive assist system to avoid such abnormalities in driving scenario as a preventive measure against occurrence of vehicle collisions, assuming that natural driving style of individual drivers is the safest style. Adaptive fuzzy system with statistics of recent fluctuations records are used to determine the driving behavior from noisy data. Another fuzzy reasoning section determines the level of abnormality in driving to notify or warn the driver so that he can pay back his full concentration in driving. Different simulated drivers with pseudo realistic styles in starting, stopping and car following are used to investigate performance of the proposed system. Empirical results show the ability of the system to recognize abnormality of drivers having different driving styles.
2021 IEEE Intelligent Vehicles Symposium (IV), 2021
The state prediction of the preceding vehicle (PV) is essential for anticipatory driving that ena... more The state prediction of the preceding vehicle (PV) is essential for anticipatory driving that enables a vehicle to take early steps for efficient driving. Such driving results in energy efficiency and traffic flow improvement but also requires repeated rigorous computations. Here we propose a Deep Neural Network (DNN) based driving scheme (DDS) that is trained to predict the PV and decide the control input of a host vehicle (HV) using two DNN models. We use the Road-Speed Profile (RSP), which provides approximated traffic speeds along the road using information from limited connected vehicles, to predict the future speed of the PV. Next, with the predicted speed of the PV, the control input is instantly determined to drive the HV efficiently. For making the optimal control decisions, the DNN is trained using a Model Predictive Control (MPC) scheme that generates the control input of the HV by minimizing a typical objective function reflecting costs related to inefficient driving. In the simulation, it is found that the proposed DDS guides the HV to change acceleration early in the optimized way by predicting the PV, particularly in changing traffic conditions, which results in a significant reduction in fuel consumption of the HV. Most remarkably, the DDS can determine the decision almost in real-time, which removes the necessity of setting a long step in the discrete-time control framework as found in MPC to match with its computation time. The proposed DDS is evaluated in some extreme cases and compared with traditional driving.
Journal of Robotics and Mechatronics, 2014
<div class=&am... more <div class=""abs_img""><img src=""[disp_template_path]/JRM/abst-image/00260005/09.jpg"" width=""300"" />Network with four intersections</div> In this paper a network-wide traffic signal control scheme in a model predictive control framework using mixed integer programming is presented. A concise model of traffic is proposed to describe a signalized road network considering conservation of traffic. In the model, the traffic of two sections that belong to a traffic signal group of a junction are represented by a single continuous variable. Therefore, the number of variables required to describe traffic in the network becomes half compared with the models that describe section wise traffic flows. The traffic signal at the junction is represented by a binary variable to express a signal state either green or red. The proposed model is transformed into a mixed logical dynamical system to describe the traffic flows in a finite horizon, and traffic signals are optimized using mixed integer linear programming (MILP) for a given performance index. The scheme simultaneously optimizes all traffic signals in a network in the context of model predictive control by successively extending or terminating a green or red signal of each junction. Consequently, traffic signal patterns with the optimal free parameters, i.e., the cycle times, the split times and the offsets, are realized. Use of the proposed concise traffic model significantly reduces the computation time of the scheme without compromising the performance as it is evaluated on a small road network and compared with a previously proposed scheme. </span>
Autonomous robots are robots which can perform desired tasks in unstructured environments without... more Autonomous robots are robots which can perform desired tasks in unstructured environments without continuous human guidance or intervention. All kinds of robots have some degree of autonomy. Different robots can be autonomous in different ways or from different angle point of view. In other word, Autonomy is the other way of describing the level of intelligence any machine has. A fully autonomous robot should have the ability to perform some basic tasks such as gaining information about the environment, work for an ...
SAE Technical Paper Series, 2015
IEEE Transactions on Vehicular Technology, 2022
Recently developed efficient driving schemes usually solve a predictive optimization problem or d... more Recently developed efficient driving schemes usually solve a predictive optimization problem or determining the vehicle control input, and at the expense of high computational cost, they improve the overall traffic flows and individual driving performances on urban roads. This paper presents a more practical technique for automated vehicles’ predictive driving by extending the existing adaptive cruise control (ACC) scheme with a look-ahead functionality. Such a look-ahead driving scheme (LDS) predicts the states of the preceding vehicle at an adaptive look-ahead time step and, with negligible computation costs, computes the vehicle control input more circumspectly for efficient driving in urban traffic. The proposed LDS is evaluated in typical urban traffic at the signalized intersections by observing the intersection utilization, flowing characteristics, and individual vehicles’ fuel efficiency. Furthermore, we also evaluate the influences of the LDS-vehicles’ penetration rates on overall traffic performances at various traffic volumes.
Electronics
Human facial emotion recognition (FER) has attracted the attention of the research community for ... more Human facial emotion recognition (FER) has attracted the attention of the research community for its promising applications. Mapping different facial expressions to the respective emotional states are the main task in FER. The classical FER consists of two major steps: feature extraction and emotion recognition. Currently, the Deep Neural Networks, especially the Convolutional Neural Network (CNN), is widely used in FER by virtue of its inherent feature extraction mechanism from images. Several works have been reported on CNN with only a few layers to resolve FER problems. However, standard shallow CNNs with straightforward learning schemes have limited feature extraction capability to capture emotion information from high-resolution images. A notable drawback of the most existing methods is that they consider only the frontal images (i.e., ignore profile views for convenience), although the profile views taken from different angles are important for a practical FER system. For deve...
A reinforcement learning algorithm is proposed that can cope with high dimensionality for a class... more A reinforcement learning algorithm is proposed that can cope with high dimensionality for a class of problems with symmetrical actions. The action selection does not need considering all the states but only needs looking at a part of the states. Moreover, every symmetrical action is related to the same kind of part of state, and thus the value function can be shared, which greatly reduces the reinforcement learning problem size. The overall learning algorithm is equivalent to the standard reinforcement learning algorithm. Simulation results and other aspects are compared with standard and other reinforcement learning algorithms. Reduction in dimensionality, much faster convergence without worsening other objectives show the effectiveness of the proposed mechanism on a high dimensional optimization problem having symmetrical actions.
2015 IEEE Intelligent Vehicles Symposium (IV), 2015
Anticipative control of vehicles is a potential approach for improving travel efficiency of indiv... more Anticipative control of vehicles is a potential approach for improving travel efficiency of individual vehicles, smoothing traffic flows on urban roads, alleviating impacts on the environment and elevating comforts of the users in various respects. This paper presents such a vehicle driving system in a model predictive control (MPC) framework to efficiently drive a vehicle on multi-lane roads. Anticipation enhances the driving intelligence and strengthens the vehicle's ability in taking advance action, e.g., lane change, speed adjustment, in a dynamically varying traffic environment. More elaborately, presuming a connected vehicle environment, the system receives the information form the surrounding vehicles and infrastructure instantly through V2X communication systems and, using dynamical models, predicts the future road-traffic states. Considering relevant constraints and a performance index, the system generates the optimal acceleration and executes lane change maneuver opti...
ABSTRACT Multiagent reinforcement learning system requires an efficient but simple way of coopera... more ABSTRACT Multiagent reinforcement learning system requires an efficient but simple way of cooperation among the autonomous agents for their successful operation in the environment. This paper presents a new approach of indirect cooperation making the agents perform the task in a competitive manner. One part of the task is performed according to a task-oriented common Q-table, which represents the information of a passive element, how the agent should work with it. The other part of the task, only related to the agent movement is done according to the agent's individual Q-table. This method converges and found very effective for dynamic environment where the goal is changed after certain episodes. Learning can be done fast and effectively even if the initialization of agents and the passive elements are random in the environment for each episode, and failure is reduced greatly.
Ieej Transactions on Electronics, Information and Systems, 2005
In this paper, a reinforcement learning method is proposed that optimizes passenger service in el... more In this paper, a reinforcement learning method is proposed that optimizes passenger service in elevator group systems. Task-oriented reinforcement learning using multiple agents is applied in the control system in allocating immediate landing calls to the elevators and operating them intelligently in attaining better service in this stochastic dynamic domain. The proposed system shows better adaptive performance in different traffic
Abstract—Sometimes a driver deviates from his natural or normal driving style due to inadequate a... more Abstract—Sometimes a driver deviates from his natural or normal driving style due to inadequate attention or faces abnormal situation caused by a number of psychological and physical factors. Such abnormalities often lead a driver to a mistake that may cause an accident. This paper presents a novel approach called driver-adaptive assist system to avoid such ab-normalities in driving scenario as a preventive measure against occurrence of vehicle collisions, assuming that natural driving style of individual drivers is the safest style. Adaptive fuzzy system with statistics of recent fluctuations records are used to determine the driving behavior from noisy data. Another fuzzy reasoning section determines the level of abnormality in driving to notify or warn the driver so that he can pay back his full concentration in driving. Different simulated drivers with pseudo realistic styles in starting, stopping and car following are used to investigate performance of the proposed system. Empi...
Multiagent reinforcement learning system requires an efficient but simple way of cooperation amon... more Multiagent reinforcement learning system requires an efficient but simple way of cooperation among the autonomous agents for their successful operation in the environment. This paper presents a new approach of indirect cooperation making the agents perform the task in a competitive manner. One part of the task is performed according to a task-oriented common Q-table, which represents the information of a passive element, how the agent should work with it. The other part of the task, only...
IET Intelligent Transport Systems, 2010
This study presents a novel concept of an ecological driver assistance system (EDAS) that may pla... more This study presents a novel concept of an ecological driver assistance system (EDAS) that may play an important role in intelligent transportation systems (ITS) in the near future. The proposed EDAS is designed to measure relevant information of instant vehicle–road–traffic utilising advanced sensing and communication technologies. Using models of vehicle dynamics and traffic flow, it anticipates future situations of the vehicle–road–traffic network, estimates fuel consumption and generates the optimal control input necessary for ecological driving. Once the optimal control input becomes available, it could be used to assist the driver through a suitable human interface. The vehicle control method is developed using model predictive control algorithm with a suitable performance index to ensure safe and fuel-efficient driving. The performance of the EDAS, in terms of speed behaviour and fuel consumption, is evaluated on the microscopic transport simulator AIMSUN NG. Comparative results are graphically illustrated and analysed to signify the prospect of the proposed EDAS in building environmentally friendly ITS.
2020 59th Annual Conference of the Society of Instrument and Control Engineers of Japan (SICE), 2020
Energy consumption of vehicles at signalized intersections is highly influenced by acceleration/d... more Energy consumption of vehicles at signalized intersections is highly influenced by acceleration/deceleration maneuvers and idling time. Existing research on fuel-efficient driving at signalized intersections mainly focused on connected and automated vehicle (CAV) technologies. This paper introduces a sustainable eco-driving strategy learning scheme from driving data of a vehicle that generates the optimal speed trajectory when approaching a signalized intersection. The goal of the proposed system is to reduce fuel consumption by advising an appropriate driving strategy near a signalized intersection. Our main contribution is that the driving performance in traffic scenarios in the existing signalized intersection can be improved without utilization of vehicle-to-vehicle (V2V) and infrastructure-to-vehicle (I2V) communications. A Gaussian process regression (GPR) model is developed that predicts intersection crossing time and crossing probability of a vehicle from its driving data. B...
2017 IEEE 20th International Conference on Intelligent Transportation Systems (ITSC), 2017
This paper addresses a partially connected vehicle environment (PCVE), where a only fraction of v... more This paper addresses a partially connected vehicle environment (PCVE), where a only fraction of vehicles are equipped with V2V communication technology, and presents a method of realizing highly anticipative driving in it. More specifically, anticipative driving refers to the predictive control of a host vehicle considering the preceding traffic conditions in an extended view but in an abstract form. For acquiring the traffic conditions of the preceding road section, a road-speed profile is proposed, which is dynamically updated using limited V2V data in a bi-level approximation framework. In the lower level of the framework, the state of the known vehicles within the communication range are predicted in a short horizon using a conditional persistence prediction model. The model parameter is determined by fitting with the experimental driving data. In the upper level, the time, position and speed of the vehicles obtained from the lower level are then mapped onto the road, which is r...
World Congress, 2011
This paper presents an ecological (eco) driving system based on prediction of the preceding vehic... more This paper presents an ecological (eco) driving system based on prediction of the preceding vehicle using model predictive control. At any measured road-traffic states it computes the optimal vehicle control input using the models of the vehicle dynamics and fuel consumption. The prediction model the preceding vehicle is formulated based on experimentally obtained driving data. The proposed system is evaluated for driving on urban roads containing traffic control signals at the intersections using the microscopic transport simulator AIMSUN. Significant improvement in fuel efficiency by introducing the model of the preceding vehicle has been confirmed from the simulation results.