Autonomous Vehicles Path Planning under Temporal Logic Specifications

TL;DR

A logic-based specification mechanism which fulfills all these requirements of path planning, originating from the necessity of collision avoidance with obstacles, respecting traffic rules, sticking to regulatory requirements, and lastly to reach the next waypoint efficiently is described.

Abstract

Path planning is an essential component of autonomous driving. A global planner is responsible for the high-level planning. It basically performs a shortest-path search on a known map, thereby defining waypoints used to control the local (low-level) planner. Local planning is a runtime verification method which is repeatedly run on the vehicle itself in real-time, so as to find the optimal short-horizon path which leads to the desired waypoint in a way which is both efficient and safe. The challenge is that the local planner has to take into account repeatedly incoming updates about the information available of the environment. In addition, it performs a complex task, as it has to take into account a large variety of requirements, originating from the necessity of collision avoidance with obstacles, respecting traffic rules, sticking to regulatory requirements, and lastly to reach the next waypoint efficiently. In this paper, we describe a logic-based specification mechanism which fulfills all these requirements.

Figures (5)

• Figure 1: Illustration of the approach. (a) Collection of human demos (in blue) and corresponding spatial GMM (red ellipses). Reproduced trajectory is in red. (b) Real-time scenario during runtime with three different constraint categories labelled in figure, the constraints breach is represented as yellow danger sign. (c) Optimized trajectory (in green) after running our algorithm.
• Figure 2: Illustration of the approach.
• Figure 3: Valet parking scenarios. The ego vehicle (in blue) must move from start to goal (depicted as diamond) while avoiding static (in red) and dynamic (in red with an arrow) obstacles. In (a), we depict the safety distance between the adjacent vehicles and the goal state. In (b), we depict the region to avoid (red cross-hatched) and region to stay (green hatched) when traffic light is red.
• Figure 4: We depict the distance to the obstacles over time before optimization (in dotted lines) and after optimization (in solid lines). A collision occurs when the distance is $0.0$. The minimum distance constraint (on the left) during time $16$ to $20$ seconds is depicted as green line.
• Figure 5: We depict the ego vehicle trajectory when traffic light is red (as dotted red line) and when it is green (as solid green line) for both before optimization (left) and after optimization (right).