Hybrid Imitation-Learning Motion Planner for Urban Driving

2024-09-04

2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC)

AI Summary

To address the trade-off between insufficient safety guarantees of learning-based planners and poor human-likeness of optimization-based planners in urban autonomous driving, this paper proposes a hybrid motion planning framework that integrates imitation learning with optimization-based control. Methodologically, we introduce the first end-to-end coupling of an MLP-based human trajectory predictor with a multi-objective nonlinear model predictive control (NMPC) refinement module, ensuring kinematic feasibility, obstacle/boundary avoidance, and explicit modeling and preservation of human driving style. Evaluated on nuPlan and Argoverse benchmarks, our approach achieves a 32% improvement in closed-loop safety rate. Furthermore, it has been successfully deployed on a real autonomous vehicle and validated over 1,000 km of urban road testing, demonstrating robustness and practical applicability.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling
• Intelligent Robots: Motion and Path Planning

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

With the release of open source datasets such as nuPlan and Argoverse, the research around learning-based planners has spread a lot in the last years. Existing systems have shown excellent capabilities in imitating the human driver behaviour, but they struggle to guarantee safe closed-loop driving. Conversely, optimization-based planners offer greater security in short-term planning scenarios. To confront this challenge, in this paper we propose a novel hybrid motion planner that integrates both learning-based and optimization-based techniques. Initially, a multilayer perceptron (MLP) generates a human-like trajectory, which is then refined by an optimization-based component. This component not only mini-mizes tracking errors but also computes a trajectory that is both kinematically feasible and collision-free with obstacles and road boundaries. Our model effectively balances safety and human-likeness, mitigating the trade-off inherent in these objectives. We validate our approach through simulation experiments and further demonstrate its efficacy by deploying it in real-world self-driving vehicles.

Problem

• Combining learning-based and optimization-based motion planning for urban driving
• Ensuring kinematically feasible and collision-free trajectories in autonomous vehicles
• Balancing safety and human-like behavior in self-driving car navigation

Innovation

• Hybrid learning and optimization motion planner
• MLP generates human-like trajectory initially
• Optimization refines for safety and feasibility