We present an iterative linear quadratic regulator(ILQR) method for trajectory tracking control of a wheeled mobile robot system.The proposed scheme involves a kinematic model linearization technique,a global trajecto...We present an iterative linear quadratic regulator(ILQR) method for trajectory tracking control of a wheeled mobile robot system.The proposed scheme involves a kinematic model linearization technique,a global trajectory generation algorithm,and trajectory tracking controller design.A lattice planner,which searches over a 3D(x,y,θ) configuration space,is adopted to generate the global trajectory.The ILQR method is used to design a local trajectory tracking controller.The effectiveness of the proposed method is demonstrated in simulation and experiment with a significantly asymmetric differential drive robot.The performance of the local controller is analyzed and compared with that of the existing linear quadratic regulator(LQR) method.According to the experiments,the new controller improves the control sequences(v,ω) iteratively and produces slightly better results.Specifically,two trajectories,'S' and '8' courses,are followed with sufficient accuracy using the proposed controller.展开更多
As a core part of an autonomous driving system,motion planning plays an important role in safe driving.However,traditional model-and rule-based methods lack the ability to learn interactively with the environment,and ...As a core part of an autonomous driving system,motion planning plays an important role in safe driving.However,traditional model-and rule-based methods lack the ability to learn interactively with the environment,and learning-based methods still have problems in terms of reliability.To overcome these problems,a hybrid motion planning framework(HMPF)is proposed to improve the performance of motion planning,which is composed of learning-based behavior planning and optimization-based trajectory planning.The behavior planning module adopts a deep reinforcement learning(DRL)algorithm,which can learn from the interaction between the ego vehicle(EV)and other human-driven vehicles(HDVs),and generate behavior decision commands based on environmental perception information.In particular,the intelligent driver model(IDM)calibrated based on real driving data is used to drive HDVs to imitate human driving behavior and interactive response,so as to simulate the bidirectional interaction between EV and HDVs.Meanwhile,trajectory planning module adopts the optimization method based on road Frenet coordinates,which can generate safe and comfortable desired trajectory while reducing the solution dimension of the problem.In addition,trajectory planning also exists as a safety hard constraint of behavior planning to ensure the feasibility of decision instruction.The experimental results demonstrate the effectiveness and feasibility of the proposed HMPF for autonomous driving motion planning in urban mixed traffic flow scenarios.展开更多
基金Project (Nos. 90920304 and 91120015) supported by the National Natural Science Foundation of China
文摘We present an iterative linear quadratic regulator(ILQR) method for trajectory tracking control of a wheeled mobile robot system.The proposed scheme involves a kinematic model linearization technique,a global trajectory generation algorithm,and trajectory tracking controller design.A lattice planner,which searches over a 3D(x,y,θ) configuration space,is adopted to generate the global trajectory.The ILQR method is used to design a local trajectory tracking controller.The effectiveness of the proposed method is demonstrated in simulation and experiment with a significantly asymmetric differential drive robot.The performance of the local controller is analyzed and compared with that of the existing linear quadratic regulator(LQR) method.According to the experiments,the new controller improves the control sequences(v,ω) iteratively and produces slightly better results.Specifically,two trajectories,'S' and '8' courses,are followed with sufficient accuracy using the proposed controller.
基金National Natural Science Foundation of China under Grants U19A2083.
文摘As a core part of an autonomous driving system,motion planning plays an important role in safe driving.However,traditional model-and rule-based methods lack the ability to learn interactively with the environment,and learning-based methods still have problems in terms of reliability.To overcome these problems,a hybrid motion planning framework(HMPF)is proposed to improve the performance of motion planning,which is composed of learning-based behavior planning and optimization-based trajectory planning.The behavior planning module adopts a deep reinforcement learning(DRL)algorithm,which can learn from the interaction between the ego vehicle(EV)and other human-driven vehicles(HDVs),and generate behavior decision commands based on environmental perception information.In particular,the intelligent driver model(IDM)calibrated based on real driving data is used to drive HDVs to imitate human driving behavior and interactive response,so as to simulate the bidirectional interaction between EV and HDVs.Meanwhile,trajectory planning module adopts the optimization method based on road Frenet coordinates,which can generate safe and comfortable desired trajectory while reducing the solution dimension of the problem.In addition,trajectory planning also exists as a safety hard constraint of behavior planning to ensure the feasibility of decision instruction.The experimental results demonstrate the effectiveness and feasibility of the proposed HMPF for autonomous driving motion planning in urban mixed traffic flow scenarios.