Transformer model guides quadrotors through static and moving obstacles
TransNav: Transformer-Based End-to-End Navigation Method for Static and Dynamic Quadrotor Obstacle Avoidance
Robotics
Summary
Flying small drones safely through crowded spaces is hard, especially when some things are still and others are moving around. The authors present a new method called TransNav that helps drones navigate around both stationary and moving obstacles by looking at the environment in a smart and efficient way. This approach uses a type of AI called a transformer to predict a series of moves directly, learning without needing expert examples. Tests in simulations and real life show it works well in complex scenes and runs fast enough to be used onboard drones.
What this means in practice
- •For drone developers: Integrate a transformer-based policy for quadrotor navigation that handles both static and moving obstacles without needing expert training data.
- •For autonomous vehicle engineers: Use environment representation and trajectory optimization techniques to improve navigation systems in cluttered and dynamic settings beyond drones.
Authors
Jianghan Zhu, Zhuo Li, Yanjun Zhang, Jian Sun
Abstract
Obstacle avoidance for autonomous quadrotors in cluttered environments remains challenging, especially when both dense static obstacles and moving obstacles are present. Existing methods are often tailored to either static or dynamic settings, which always limit their adaptability across diverse environments. This paper presents TransNav, a transformer-based end-to-end navigation method for quadrotor obstacle avoidance in both static and dynamic environments. TransNav employs a patch-based point-cloud encoding scheme that combines local feature extraction with global contextual modeling, enabling compact yet informative environment representation. The method integrates perception and navigation in a single architecture and directly predicts a finite-horizon action sequence, from which the first action is executed. Moreover, we develop an unsupervised learning strategy by incorporating a lightweight point-mass dynamics model, which directly optimizes trajectory-level costs over the predicted sequence through numerical gradient propagation. Thus, the policy can be efficiently trained without expert demonstrations. Simulation and real-world experiments show that the proposed method achieves strong performance in dense static environments and competitive obstacle-avoidance capability in dynamic scenarios, while maintaining low inference cost for onboard deployment.