Learning and Adapting Behavior of Autonomous Vehicles through Inverse Reinforcement Learning

Trauth, Rainer; Kaufeld, Marc; Geisslinger, Maximilian; Betz, Johannes

doi:10.1109/iv55152.2023.10186668

Cited by 3 publications

(2 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Which weightings and settings are the best still needs to be evaluated. How to adjust cost parameters in general has been explored in other work [48], [49]. The setting of these parameters generally depends highly on the system architecture of the analytical models.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Toward Safer Autonomous Vehicles: Occlusion-Aware Trajectory Planning to Minimize Risky Behavior

Trauth,

Moller,

Betz

2023

IEEE Open J. Intell. Transp. Syst.

Self Cite

View full text Add to dashboard Cite

Autonomous vehicles face numerous challenges to ensure safe operation in unpredictable and hazardous conditions. The autonomous driving environment is characterized by high uncertainty, especially in occluded areas with limited information about the surrounding obstacles. This work aims to provide a trajectory planner to solve these unsafe environments. The work proposes an approach combining a visibility model, contextual environmental information, and behavioral planning algorithms to predict the likelihood of occlusions and collision probabilities. Ultimately, this allows us to estimate the potential harm from collisions with pedestrians in occluded situations. The primary goal of our proposed approach is to minimize the risk of hitting pedestrians and to establish a predefined, adjustable maximum level of harm. We show several practical applications for informing a sampling-based trajectory planner about occluded areas to increase safety. In addition, to respond to possible high-risk situations, we introduce an adjustable threshold that governs the vehicle's speed when encountering uncertain situations and strategies to maximize the vehicle's visible area. In implementing our novel methodology, we analyzed several realworld scenarios in a simulation environment. Our results indicate that combining occlusion-aware trajectory planning algorithms and harm estimation significantly influences vehicle driving behavior, especially in risky situations. The code used in this research is publicly available as open-source software and can be accessed at the following link: https://github.com/TUM-AVS/Frenetix-Motion-Planner.

show abstract

Section: Discussionmentioning

confidence: 99%

“…As with all analytical models, the weighting factor ω must be chosen carefully. Previous work has investigated how to find weights in specific domains [48] or how to derive weights with supervised learning from, for example, human data [49]. The weighting factor ω linearly affects the trajectory selection decision.…”

Section: Harm Evaluationmentioning

confidence: 99%

Toward Safer Autonomous Vehicles: Occlusion-Aware Trajectory Planning to Minimize Risky Behavior

Trauth,

Moller,

Betz

2023

IEEE Open J. Intell. Transp. Syst.

Self Cite

View full text Add to dashboard Cite

show abstract

Simplifying Sim-to-Real Transfer in Autonomous Driving: Coupling Autoware with the CommonRoad Motion Planning Framework

Würsching,

Mascetta,

Lin

et al. 2024

2024 IEEE Intelligent Vehicles Symposium (IV)

View full text Add to dashboard Cite

Validating motion planning algorithms for autonomous vehicles on a real system is essential to improve their safety in the real world. Open-source initiatives, such as Autoware, provide a deployable software stack for real vehicles. However, such driving stacks have a high entry barrier, so that integrating new algorithms is tedious. Especially new research results are thus mostly evaluated only in simulation, e.g., within the CommonRoad benchmark suite. To address this problem, we present CR2AW, a publicly available interface between the CommonRoad framework and Autoware. CR2AW significantly simplifies the sim-to-real transfer of motion planning research, by allowing users to easily integrate their CommonRoad planning modules into Autoware. Our experiments both in simulation and on our research vehicle showcase the usefulness of CR2AW.

show abstract

A Reinforcement Learning-Boosted Motion Planning Framework: Comprehensive Generalization Performance in Autonomous Driving

Trauth,

Hobmeier,

Betz

2024

2024 IEEE Intelligent Vehicles Symposium (IV)

View full text Add to dashboard Cite

Learning and Adapting Behavior of Autonomous Vehicles through Inverse Reinforcement Learning

Cited by 3 publications

References 26 publications

Toward Safer Autonomous Vehicles: Occlusion-Aware Trajectory Planning to Minimize Risky Behavior

Toward Safer Autonomous Vehicles: Occlusion-Aware Trajectory Planning to Minimize Risky Behavior

Simplifying Sim-to-Real Transfer in Autonomous Driving: Coupling Autoware with the CommonRoad Motion Planning Framework

A Reinforcement Learning-Boosted Motion Planning Framework: Comprehensive Generalization Performance in Autonomous Driving

Contact Info

Product

Resources

About