Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Enhancing System Self-Awareness and Trust of AI: A Case Study in Trajectory Prediction and Planning

Lars Ullrich, Zurab Mujirishvili, Knut Graichen

TL;DR

The proposed TrustMHE represents a complementary approach, independent of the underlying AI systems, that combines AI-driven out-of-distribution detection with control-driven moving horizon estimation (MHE) to enable not only detection and monitoring, but also intervention.

Abstract

In the trajectory planning of automated driving, data-driven statistical artificial intelligence (AI) methods are increasingly established for predicting the emergent behavior of other road users. While these methods achieve exceptional performance in defined datasets, they usually rely on the independent and identically distributed (i.i.d.) assumption and thus tend to be vulnerable to distribution shifts that occur in the real world. In addition, these methods lack explainability due to their black box nature, which poses further challenges in terms of the approval process and social trustworthiness. Therefore, in order to use the capabilities of data-driven statistical AI methods in a reliable and trustworthy manner, the concept of TrustMHE is introduced and investigated in this paper. TrustMHE represents a complementary approach, independent of the underlying AI systems, that combines AI-driven out-of-distribution detection with control-driven moving horizon estimation (MHE) to enable not only detection and monitoring, but also intervention. The effectiveness of the proposed TrustMHE is evaluated and proven in three simulation scenarios.

Enhancing System Self-Awareness and Trust of AI: A Case Study in Trajectory Prediction and Planning

TL;DR

The proposed TrustMHE represents a complementary approach, independent of the underlying AI systems, that combines AI-driven out-of-distribution detection with control-driven moving horizon estimation (MHE) to enable not only detection and monitoring, but also intervention.

Abstract

In the trajectory planning of automated driving, data-driven statistical artificial intelligence (AI) methods are increasingly established for predicting the emergent behavior of other road users. While these methods achieve exceptional performance in defined datasets, they usually rely on the independent and identically distributed (i.i.d.) assumption and thus tend to be vulnerable to distribution shifts that occur in the real world. In addition, these methods lack explainability due to their black box nature, which poses further challenges in terms of the approval process and social trustworthiness. Therefore, in order to use the capabilities of data-driven statistical AI methods in a reliable and trustworthy manner, the concept of TrustMHE is introduced and investigated in this paper. TrustMHE represents a complementary approach, independent of the underlying AI systems, that combines AI-driven out-of-distribution detection with control-driven moving horizon estimation (MHE) to enable not only detection and monitoring, but also intervention. The effectiveness of the proposed TrustMHE is evaluated and proven in three simulation scenarios.

Paper Structure

This paper contains 13 sections, 18 equations, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Trajectory Prediction & Planning
  3. Motion Transformer-Based Trajectory Prediction
  4. Model Predictive Path Integral-Based Trajectory Planning
  5. Integration & Implementation
  6. TrustMHE
  7. Core Principles
  8. Methodological Concept
  9. Implementation & Integration
  10. Empirical Evaluation
  11. Experimental Setup
  12. Results
  13. Discussion & Conclusion

Figures (3)

  • Figure 1: Simplified representation of MTR predictor ullrich2024transfer.
  • Figure 2: Illustration of the closed-loop experimental setup.
  • Figure 3: Evaluation of TrustMHE across different experimental setting variations. Overall TrustMHE "Disabled" is compared against five slightly different TrustMHE "Enabled" settings. In particular, five different horizons $T_{\mathrm{est}}=\{1,3,5,15,30\}$ settings are considered and separately illustrated.