Motion Planning for Autonomous Driving with Real Traffic Data Validation

Accurate trajectory prediction of surrounding road users is the fundamental input for motion planning,which enables safe autonomous driving on public roads.In this paper,a safe motion planning approach is proposed based on the deep learning-based trajectory prediction method.To begin with,a trajectory prediction model is established based on the graph neural network(GNN)that is trained utilizing the INTERACTION dataset.Then,the validated trajectory prediction model is used to predict the future trajectories of surrounding road users,including pedestrians and vehicles.In addition,a GNN prediction model-enabled motion planner is developed based on the model predictive control technique.Furthermore,two driving scenarios are extracted from the INTERACTION dataset to validate and evaluate the effectiveness of the proposed motion planning approach,i.e.,merging and roundabout scenarios.The results demonstrate that the proposed method can lower the risk and improve driving safety compared with the baseline method.

MPC-based Motion Planning and Control Enables Smarter and Safer Autonomous Marine Vehicles:Perspectives and a Tutorial Survey

Autonomous marine vehicles(AMVs)have received considerable attention in the past few decades,mainly because they play essential roles in broad marine applications such as environmental monitoring and resource exploration.Recent advances in the field of communication technologies,perception capability,computational power and advanced optimization algorithms have stimulated new interest in the development of AMVs.In order to deploy the constrained AMVs in the complex dynamic maritime environment,it is crucial to enhance the guidance and control capabilities through effective and practical planning,and control algorithms.Model predictive control(MPC)has been exceptionally successful in different fields due to its ability to systematically handle constraints while optimizing control performance.This paper aims to provide a review of recent progress in the context of motion planning and control for AMVs from the perceptive of MPC.Finally,future research trends and directions in this substantial research area of AMVs are highlighted.

Inverted Modelling:An Effective Way to Support Motion Planning of Legged Mobile Robots

This paper presents an effective way to support motion planning of legged mobile robots—Inverted Modelling,based on the equivalent metamorphic mechanism concept.The difference from the previous research is that we herein invert the equivalent parallel mechanism.Assuming the leg mechanisms are hybrid links,the body of robot being considered as fixed platform,and ground as moving platform.The motion performance is transformed and measured in the body frame.Terrain and joint limits are used as input parameters to the model,resulting in the representation which is independent of terrains and particular poses in Inverted Modelling.Hence,it can universally be applied to any kind of legged robots as global motion performance framework.Several performance measurements using Inverted Modelling are presented and used in motion performance evaluation.According to the requirements of actual work like motion continuity and stability,motion planning of legged robot can be achieved using different measurements on different terrains.Two cases studies present the simulations of quadruped and hexapod robots walking on rugged roads.The results verify the correctness and effectiveness of the proposed method.

A review of mobile robot motion planning methods:from classical motion planning workflows to reinforcement learning-based architectures

Motion planning is critical to realize the autonomous operation of mobile robots.As the complexity and randomness of robot application scenarios increase,the planning capability of the classical hierarchical motion planners is challenged.With the development of machine learning,the deep reinforcement learning(DRL)-based motion planner has gradually become a research hotspot due to its several advantageous feature.The DRL-based motion planner is model-free and does not rely on the prior structured map.Most importantly,the DRL-based motion planner achieves the unification of the global planner and the local planner.In this paper,we provide a systematic review of various motion planning methods.Firstly,we summarize the representative and state-of-the-art works for each submodule of the classical motion planning architecture and analyze their performance features.Then,we concentrate on summarizing reinforcement learning(RL)-based motion planning approaches,including motion planners combined with RL improvements,map-free RL-based motion planners,and multi-robot cooperative planning methods.Finally,we analyze the urgent challenges faced by these mainstream RLbased motion planners in detail,review some state-of-the-art works for these issues,and propose suggestions for future research.

Rollover Prevention and Motion Planning for an Intelligent Heavy Truck

It is very necessary for an intelligent heavy truck to have the ability to prevent rollover independently.However,it was rarely considered in intelligent vehicle motion planning.To improve rollover stability,a motion planning strategy with autonomous anti rollover ability for an intelligent heavy truck is put forward in this paper.Considering the influence of unsprung mass in the front axle and the rear axle and the body roll stiffness on vehicle rollover stability,a rollover dynamics model is built for the intelligent heavy truck.From the model,a novel rollover index is derived to evaluate vehicle rollover risk accurately,and a model predictive control algorithm is applicated to design the motion planning strategy for the intelligent heavy truck,which integrates the vehicle rollover stability,the artificial potential field for the obstacle avoidance,the path tracking and vehicle dynamics constrains.Then,the optimal path is obtained to meet the requirements that the intelligent heavy truck can avoid obstacles and drive stably without rollover.In addition,three typical scenarios are designed to numerically simulate the dynamic performance of the intelligent heavy truck.The results show that the proposed motion planning strategy can avoid collisions and improve vehicle rollover stability effectively even under the worst driving scenarios.

Post-Impact Motion Planning and Tracking Control for Autonomous Vehicles

There is an increasing awareness of the need to reduce traffic accidents and fatality due to vehicle collision.Post-impact hazards can be more serious as the driver may fail to maintain effective control after collisions.To avoid subsequent crash events and to stabilize the vehicle,this paper proposes a post-impact motion planning and stability control method for autonomous vehicles.An enabling motion planning method is proposed for post-impact situations by combining the polynomial curve and artificial potential field while considering obstacle avoidance.A hierarchical controller that consists of an upper and a lower controller is then developed to track the planned motion.In the upper controller,a time-varying linear quadratic regulator is presented to calculate the desired generalized forces.In the lower controller,a nonlinear-optimization-based torque allocation algorithm is proposed to optimally coordinate the actuators to realize the desired generalized forces.The proposed scheme is verified under comprehensive driving scenarios through hardware-in-loop tests.

Driving Environment Uncertainty-Aware Motion Planning for Autonomous Vehicles

Autonomous vehicles require safe motion planning in uncertain environments,which are largely caused by surrounding vehicles.In this paper,a driving environment uncertainty-aware motion planning framework is proposed to lower the risk of position uncertainty of surrounding vehicles with considering the risk of rollover.First,a 4-degree of freedom vehicle dynamics model,and a rollover risk index are introduced.Besides,the uncertainty of surrounding vehicles’position is processed and propagated based on the Extended Kalman Filter method.Then,the uncertainty potential field is established to handle the position uncertainty of autonomous vehicles.In addition,the model predictive controller is designed as the motion planning framework which accounts for the rollover risk,the position uncertainty of the surrounding vehicles,and vehicle dynamic constraints of autonomous vehicles.Furthermore,two edge cases,the cut-in scenario,and merging scenario are designed.Finally,the safety,effectiveness,and real-time performance of the proposed motion planning framework are demonstrated by employing a hardware-in-the-loop experiment bench.

Genetic algorithm-fuzzy based dynamic motion planning approach for a mobile robot

Presents the mobile robots dynamic motion planning problem with a task to find an obstacle free route that requires minimum travel time from the start point to the destination point in a changing environment, due to the obstacle’s moving. An Genetic Algorithm fuzzy (GA Fuzzy) based optimal approach proposed to find any obstacle free path and the GA used to select the optimal one, points out that using this learned knowledge off line, a mobile robot can navigate to its goal point when it faces new scenario on line. Concludes with the optimal rule base given and the simulation results showing its effectiveness.

A motion planning algorithm for the feed support system of FAST

The paper relates to a motion planning algorithm for the feed support system of the Five-hundred-meter Aperture Spherical radio Telescope(FAST).To enhance the stability of the feed support system,the start/termination planning segments are adopted with an acceleration and deceleration section.The source switching planning adopts a combination of a line segment and focal segment to realize stable control of the feed support system.Besides,during the observation trajectory,a transition segment which is not used for observation data is planned with a required time.Through an example simulation,a smooth change is realized via the motion planning algorithm and presented in this paper.

Motion Planning System for Bin Picking Using 3-D Point Cloud

In this paper, we propose a motion planning system for bin picking using 3-D point cloud. The situation that the objects are put miscellaneously like the inside in a house is assumed. In the home, the equipment which makes an object stand in line doesn’t exist. Therefore the motion planning system which considered a collision problem becomes important. In this paper, Information on the objects is measured by a laser range finder (LRF). The information is used as 3-D point cloud, and the objects are recognized by model-base. We propose search method of a grasping point for two-fingered robotic hand, and propose search method of a path to approach the grasping point without colliding with other objects.

Motion Planning for Minimizing Base Disturbance of Space Manipulators

A simulated annealing particle swarm optimization(PSO)algorithm is utilized in this paper to plan the motion of space manipulators with minimized base disturbance,not only its attitude but also its position.Since space manipulators must meet the Law of Momentum Conservation,any motion of manipulators will disturb the spacecraft which is free-float in the space environment.This paper tries to limit this effect to the minimum.First,this paper establishes the mathematical model for space manipulators by generalized Jacobian matrix(GJM) and analyzes its inverse kinematics by Theory of Screws.Second,a polynomial function of seventh degree is used to parameterize the joint motion and quaternion representation is also used to represent attitude of spacecraft.Moreover,this paper designs a proper objective function and depicts this algorithm detailedly and clearly.Finally,the results of numerical simulation are verified by the proposed algorithm.

Kinematics Modeling and Motion Planning of Tendon Driven Continuum Manipulators

Continuum manipulators have important applications in the human–machine interaction fields.The kinematics,dynamics,and control issues of the continuum manipulators are rather different from a conventional rigid-link manipulator.By the aid of Lie groups theory and exponential coordinate representations,the kinematics of the continuum manipulators with piecewise constant curvatures and actuated by tendons is investigated in this paper.On the basis of differential kinematics analysis,the complete Jacobian of the continuum manipulators is derived analytically,and then a new motion planning approach,named as“dynamic coordination method”is presented for the multisegments continuum manipulators,which is a class of superredundant manipulators.The novel motion planning approach is featured by some appealing properties,and the feasibility of the modeling and the motion planning method is demonstrated by some numerical simulations.

Motion Planning Based Coordinated Control for Hydraulic Excavators

Hydraulic excavator is one type of the most widely applied construction equipment for various applications mainly because of its versatility and mobility.Among the tasks performed by a hydraulic excavator, repeatable level digging or flat surface finishing may take a large percentage.Using automated functions to perform such repeatable and tedious jobs will not only greatly increase the overall productivity but more importantly also improve the operation safety.For the purpose of investigating the technology without loss of generality, this research is conducted to create a coordinate control method for the boom, arm and bucket cylinders on a hydraulic excavator to perform accurate and effective works.On the basis of the kinematic analysis of the excavator linkage system, the tip trajectory of the end-effector can be determined in terms of three hydraulic cylinders coordinated motion with a visualized method.The coordination of those hydraulic cylinders is realized by controlling three electro-hydraulic proportional valves coordinately.Therefore, the complex control algorithm of a hydraulic excavator can be simplified into coordinated motion control of three individual systems.This coordinate control algorithm was validated on a wheeled hydraulic excavator, and the validation results indicated that this developed control method could satisfactorily accomplish the auto-digging function for level digging or flat surface finishing.

Motion Planning Algorithms of Redundant Manipulators Based on Self-motion Manifolds

The current motion planning approaches for redundant manipulators mainly includes two categories: improved gradient-projection method and some other efficiency numerical methods. The former is excessively sensitive to parameters, which makes adjustment difficult; and the latter treats the motion planning as general task by ignoring the particularity, which has good universal property but reduces the solving speed for on-line real-time planning. In this paper, a novel stepwise solution based on self-motion manifold is proposed for motion planning of redundant manipulators, namely, the chief tasks and secondary tasks are implemented step by step. Firstly, the posture tracking of end-effector is achieved accurately by employing the non-redundant joint. Secondly, the end-effector is set to keep stationary. Finally, self-motion of manipulator is realized via additional work on the gradient of redundant joint displacement. To verify this solution, experiments of round obstacle avoiding are carried out via the planar 3 degree-of-freedom manipulator. And the experimental results indicate that this motion planning algorithm can effectively achieve obstacle avoiding and posture tracking of the end-effector. Compared with traditional gradient projection method, this approach can accelerate the problem-solving process, and is more applicable to obstacle avoiding and other additional work in displacement level.

Parallel Planning:A New Motion Planning Framework for Autonomous Driving

Motion planning is one of the most significant technologies for autonomous driving. To make motion planning models able to learn from the environment and to deal with emergency situations, a new motion planning framework called as"parallel planning" is proposed in this paper. In order to generate sufficient and various training samples, artificial traffic scenes are firstly constructed based on the knowledge from the reality.A deep planning model which combines a convolutional neural network(CNN) with the Long Short-Term Memory module(LSTM) is developed to make planning decisions in an end-toend mode. This model can learn from both real and artificial traffic scenes and imitate the driving style of human drivers.Moreover, a parallel deep reinforcement learning approach is also presented to improve the robustness of planning model and reduce the error rate. To handle emergency situations, a hybrid generative model including a variational auto-encoder(VAE) and a generative adversarial network(GAN) is utilized to learn from virtual emergencies generated in artificial traffic scenes. While an autonomous vehicle is moving, the hybrid generative model generates multiple video clips in parallel, which correspond to different potential emergency scenarios. Simultaneously, the deep planning model makes planning decisions for both virtual and current real scenes. The final planning decision is determined by analysis of real observations. Leveraging the parallel planning approach, the planner is able to make rational decisions without heavy calculation burden when an emergency occurs.

Optimal Motion Planning for Differentially Flat Underactuated Mechanical Systems

Underactuated mechanical system has less independent inputs than the degrees of freedom(DOF) of the mechanism. The energy efficiency of this class of mechanical systems is an essential problem in practice. On the basis of the sufficient and necessary condition that concludes a single input nonlinear system is differentially flat, it is shown that the flat output of the single input underactuated mechanical system can be obtained by finding a smooth output function such that the relative degree of the system equals to the dimension of the state space. If the flat output of the underactuated system can be solved explicitly, and by constructing a smooth curve with satisfying given boundary conditions in flat output space, an energy efficiency optimization method is proposed for the motion planning of the differentially flat underactuated mechanical systems. The inertia wheel pendulum is used to verify the proposed optimization method, and some numerical simulations show that the presented optimal motion planning method can efficaciously reduce the energy cost for given control tasks.

A sensor-based motion planner: Situated-Bug

Propose a new sensor-based motion planning approach of Situated-Bug, which is composed of goal-oriented behavior, boundary following behavior and goal-oriented obstacle avoidance behavior, which are based on fuzzy control. The situated-Bug selects its behaviors according to robot orientation, instead of positions and hit points like other Bug algorithms, and its convergence proves robust to sensor noise, and it can guide the robots running for long rang traverse. At the same time, the design of the Situated-Bug is presented. Simulation results show that the approach is effective and practical.

Efficient Configuration Space Construction and Optimization for Motion Planning

The configuration space is a fundamental conc ept that is widely used in algorithmic robotics. Many applications in robotics, computer-aided design, and related areas can be reduced to computational problems in terms of configuration spaces. In this paper, we survey some of our recent work on solving two important challenges related to configuration spaces:(1) how to efficiently compute an approximate representation of high-dimensional configuration spaces; and(2) how to efficiently perform geometric proximity and motion planning queries in high-dimensional configuration spaces. We present new configuration space construction algorithms based on machine learning and geometric approximation techniques. These algorithms perform collision queries on many configuration samples. The collision query results are used to compute an approximate representation for the configuration space, which quickly converges to the exact configuration space. We also present parallel GPU-based algorithms to accelerate the performance of optimization and search computations in configuration spaces. In particular, we design efficient GPU-based parallel k-nearest neighbor and parallel collision detection algorithms and use these algorithms to accelerate motion planning.

Motion Planning of Humanoid Robot for Obstacle Negotiation

The motion planning for obstacle negotiation by humanoid robot BHR-2 through stepping over or stepping on/off the wide and flat obstacle at known locations is presented. In the trajectory generation method,first the constraints of the foot motion parameters which include obstacle dimensions and the distance of obstacle from the humanoid robot is formulated. By varying the values of the constraint parameters,different types of foot motion for different obstacles can be produced. In this method,first the foot trajectory is generated,and then the waist trajectory is computed by using cubic spline interpolation without first calculating the zero moment point (ZMP) trajectory. The dynamic stability during the execution of stepping over and stepping on/off trajectories are ensured by incorporating the ZMP criterion. The effectiveness of the proposed method is confirmed by simulations and experiments on humanoid robot BHR-2.