Implicit Maximum Likelihood Estimation for Real-Time Generative Model Predictive Control
Grayson Lee, Minh Bui, Shuzi Zhou, Yankai Li, Mo Chen, Ke Li
AI summary
Problem
Diffusion-based generative planners for trajectory optimization suffer from slow inference speeds due to iterative denoising, making them unsuitable for real-time, closed-loop model predictive control in dynamic environments.
Approach
The authors adapt Implicit Maximum Likelihood Estimation to generate diverse trajectory candidates in a single forward pass, using reward-weighted training to bias generation toward high-return paths, integrated into sampling-based MPC frameworks.
Key results
- Competitive performance on D4RL MuJoCo locomotion benchmarks
- Two orders of magnitude faster sampling frequency than diffusion models
- Real-time closed-loop navigation for mobile robots among pedestrians
- Effective reward-weighted training for mixed-quality offline data
Why it matters
Provides a computationally efficient alternative to diffusion models for real-time robotic planning and control, bridging the gap between high-quality generative modeling and latency-sensitive applications.
Abstract
Diffusion-based models have recently shown strong performance in trajectory planning, as they are capable of capturing diverse, multimodal distributions of complex behaviors. A key limitation of these models is their slow inference speed, which results from the iterative denoising process. This makes them less suitable for real-time applications such as closed-loop model predictive control (MPC), where plans must be generated quickly and adapted continuously to a changing environment. In this paper, we investigate Implicit Maximum Likelihood Estimation (IMLE) as an alternative generative modeling approach for planning. IMLE offers strong mode coverage while enabling inference that is two orders of magnitude faster, making it particularly well suited for real-time MPC tasks. Our results demonstrate that IMLE achieves competitive performance on standard offline reinforcement learning benchmarks compared to the standard diffusion-based planner, while substantially improving planning speed in both open-loop and closed-loop settings. We further validate IMLE in a closed-loop human navigation scenario, operating in real-time, demonstrating how it enables rapid and adaptive plan generation in dynamic environments. Real-world videos and code are available at https://gmpc-imle.github.io/.