Fail-Safe Motion Planning for Online Verification of Autonomous Vehicles Using Convex Optimization

Abstract

This is the accepted version of a paper published in IEEE Transactions on robotics. This paper has been peer-reviewed but does not include the final publisher proof-corrections or journal pagination.

Online Trajectory Replanning for Sudden Environmental Changes During Automated Parking: A Parallel Stitching Method

“…Concretely, an emerging obstacle is regarded as a failure if it blocks the original trajectory. Assuming that such failures may happen in the near future, a fail-safe planner generates evasive trajectories as precautions [18]. However, this method is more suitable for on-road cruising scenarios because the underlying failures in an unstructured parking scenario are difficult to enumerate.…”

Risk-aware Motion Planning for Autonomous Vehicles with Safety Specifications

“…This technique can be used to determine if the AV is able to reach any unsafe set of states [14], [15] or to rigorously predict the possible occupied space of other traffic participants [16]- [18]. Reachability analysis is well suited to verify the safety of the AV during its operation as shown in [19]- [22]. However, in these approaches it is usually difficult to estimate whether the ego vehicle is close to a violation.…”

“…Invariant sets are sets of safe states which allow the AV to remain safe [23]. With these sets, one is able to detect when the AV performs trajectories that are close to violating safety constraints, simply by checking the distance to the boundary of the set [19]. Recent approaches show that invariant sets can be computed using under-approximations [24], reachability analysis [25], control barrier functions [26], [27] or classical control techniques [28], [29].…”

Safe Data-Driven Contact-Rich Manipulation

“…Similarly, invariant sets (IS) contain states which allow the system to remain safe [18]- [21], i.e., there exists at least one safe action for each state within an IS. Unfortunately, CBFs and IS generally require an analytical model of the system to compute the set of safe states.…”