A Review of Motion Planning for Highway Autonomous Driving

I. INTRODUCTION

II. CONSIDERATIONS FOR HIGHWAY MOTION PLANNING

A. Terminology

B. Motion Planning Scheme

The core functions addressed in this review are decision making (iii), generation (iv), and deformation (v), which we call the motion planning scope.

C. Specificities of Highway Driving

 lanedivided roads featuring unidirectional flow 

a dynamic speed over 60km/h.

The road shape is made of straight lines, clothoids, and circles with small curvature.

only motorized vehicles, which adhere to the same driving rules.

Obstacle’s behavior prediction is also limited to one-direction, two-lane changes – right or left – and to accelerate, maintain speed, or brake.

distinguish eight nominal nonexclusive situations：Lane keeping、Car following、Lane changing、Lateral-most lane changing、Passing、Overtaking、Merging、Highway toll

D. Constraints on Highway Driving

hard constraints：respect the driving rules and avoid collision

soft constraints：the minimization of time, distance, or energy consumption, and maximization of comfort

kinematic restrictions of the vehicle（自由度）

E. What Is at Stake in this Paper

classify methods from a practical perspective, i.e.

analyze how the algorithms work well, their advantages and drawbacks, considering simulations or experiments.

III. STATE OF THE ART

A. Taxonomy Description（分类方法描述）

The type of output、The space-time property of the algorithm、The mathematical domains（ geometric, heuristic, logic, cognitive, biomimetic）

B. Algorithm Classification

1) Space Configuration

2) Pathfinding Algorithms

3) Attractive and Repulsive Forces

4) Parametric and Semi-parametric Curves

5) Numerical Optimization

Linear Programming (LP)、Quadratic Programming optimization (QP)、Model Predictive Control (MPC)、Dynamic Programming (DP)

6) Artificial Intelligence

IV. COMPARISON TABLE FOR HIGHWAY APPLICATIONS

V. FUTURE RESEARCH DIRECTIONS

A. Data Management  

1) Data Uncertainty and Trustworthiness

2) Data latency and unsynchronized

B. Adaptive Mobility

safety, eco-driving, perception compensation, route context, and serviceoriented vehicles.

C. Cooperative Planning

individual, collective, and shared.

D. Validation and Evaluation

Transition Stability（转换平稳性）Ethics（伦理）Algorithm Relaxation（硬约束可能变得危险）

VI. CONCLUSION

encourage readers’ reflections on the practical implementation of these algorithms