Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Intention-Aware Control Based on Belief-Space Specifications and Stochastic Expansion

Zengjie Zhang, Zhiyong Sun, Sofie Haesaert

TL;DR

This work provides a novel solution for the risk-aware control of interactive vehicles with formal safety guarantees by solving an intention-aware control problem incorporating epistemic uncertainties of the opponent vehicles and model their intentions as discrete-valued random variables.

Abstract

This paper develops a correct-by-design controller for an autonomous vehicle interacting with opponent vehicles with unknown intentions. We define an intention-aware control problem incorporating epistemic uncertainties of the opponent vehicles and model their intentions as discrete-valued random variables. Then, we focus on a control objective specified as belief-space temporal logic specifications. From this stochastic control problem, we derive a sound deterministic control problem using stochastic expansion and solve it using shrinking-horizon model predictive control. The solved intention-aware controller allows a vehicle to adjust its behaviors according to its opponents' intentions. It ensures provable safety by restricting the probabilistic risk under a desired level. We show with experimental studies that the proposed method ensures strict limitation of risk probabilities, validating its efficacy in autonomous driving cases. This work provides a novel solution for the risk-aware control of interactive vehicles with formal safety guarantees.

Intention-Aware Control Based on Belief-Space Specifications and Stochastic Expansion

TL;DR

This work provides a novel solution for the risk-aware control of interactive vehicles with formal safety guarantees by solving an intention-aware control problem incorporating epistemic uncertainties of the opponent vehicles and model their intentions as discrete-valued random variables.

Abstract

This paper develops a correct-by-design controller for an autonomous vehicle interacting with opponent vehicles with unknown intentions. We define an intention-aware control problem incorporating epistemic uncertainties of the opponent vehicles and model their intentions as discrete-valued random variables. Then, we focus on a control objective specified as belief-space temporal logic specifications. From this stochastic control problem, we derive a sound deterministic control problem using stochastic expansion and solve it using shrinking-horizon model predictive control. The solved intention-aware controller allows a vehicle to adjust its behaviors according to its opponents' intentions. It ensures provable safety by restricting the probabilistic risk under a desired level. We show with experimental studies that the proposed method ensures strict limitation of risk probabilities, validating its efficacy in autonomous driving cases. This work provides a novel solution for the risk-aware control of interactive vehicles with formal safety guarantees.
Paper Structure (20 sections, 32 equations, 8 figures)

This paper contains 20 sections, 32 equations, 8 figures.

Table of Contents

  1. Introduction
  2. Problem Statement
  3. Description of Intention-Aware Control
  4. Problem Formulation
  5. Mathematical Models
  6. Modeling Traffic Participants
  7. Random Intentions with Epistemic Uncertainty
  8. Quantifying Risk Using Probabilistic Constraints
  9. Describing Tasks Using Belief-Space Specifications
  10. Tractable Control Solution Based on PCE
  11. Experimental Studies
  12. Case I: Overtaking
  13. Case II: Intersection
  14. DISCUSSION
  15. CONCLUSIONS
  16. ...and 5 more sections

Figures (8)

  • Figure 1: An intersection case where an plans a collision-free left turn, aiming at a safe trajectory for the complex modeling-intention uncertainties of the and pedestrians. Understanding the intentions of the and the pedestrians helps the make safe decisions.
  • Figure 2: The intention-aware control framework.
  • Figure 3: The overtaking case, where an in the fast lane should switch to the slow lane and overtake an within a finite time (the dashed arrow). The area in front of the bounded by two dotted arrows denotes the joint epistemic uncertainty at a certain risk level.
  • Figure 4: The trajectory of the and sampled trajectories of the .
  • Figure 5: The computation time per step (Case I).
  • ...and 3 more figures

Theorems & Definitions (3)

  • Definition 1
  • Remark 1: Robustness
  • Remark 2: Computational Complexity