DTG: Diffusion-based Trajectory Generation for Mapless Global Navigation

We present a novel end-to-end diffusion-based trajectory generation method, DTG, for mapless global navigation in challenging outdoor scenarios with occlusions and unstructured off-road features like grass, buildings, bushes, etc. Given a distant goal, our approach computes a trajectory that satisfies the following goals: (1) minimize the travel distance to the goal; (2) maximize the traversability by choosing paths that do not lie in undesirable areas. Specifically, we present a novel Conditional RNN(CRNN) for diffusion models to efficiently generate trajectories. Furthermore, we propose an adaptive training method that ensures that the diffusion model generates more traversable trajectories. We evaluate our methods in various outdoor scenes and compare the performance with other global navigation algorithms on a Husky robot. In practice, we observe at least a 15% improvement in traveling distance and around a 7% improvement in traversability. Video and Code: https://github.com/jingGM/DTG.git.

Authors

• Jing Liang 0006
• Amirreza Payandeh
• Daeun Song
• Xuesu Xiao
• Dinesh Manocha

Keywords

• Training
• Codes
• Navigation
• Buildings
• Diffusion models
• Trajectory
• Intelligent robots
• Trajectory Generation
• Diffusion Model
• Training Adaptations
• Neural Network
• Terrain
• Gaussian Noise
• Generative Adversarial Networks
• Shortest Distance
• Path Planning
• Inference Time
• Learning-based Approaches
• Real-world Experiments
• Trajectory Optimization
• Distance Ratio
• Trajectory Length
• Navigation Task
• Robot Navigation
• Large-scale Environments
• Vector C
• Ground Truth Trajectory
• Robots In Environments
• Odometer
• Evaluation Dataset
• Topographic Maps