Think Fast: Real-Time Kinodynamic Belief-Space Planning for Projectile Interception

TL;DR

This paper presents a real-time kinodynamic belief-space planning method for intercepting fast-moving objects with noisy sensor data, using a tree structure in state-time space and adaptive filtering on a 6 DOF robot arm.

Contribution

It introduces a novel tree-based planning framework that handles belief updates and goal transitions in real-time for projectile interception tasks.

Findings

Successfully intercepted targets with a 6 DOF robot arm.

Real-time updates improved interception accuracy.

Adaptive filtering enhanced target tracking robustness.

Abstract

Intercepting fast moving objects, by its very nature, is challenging because of its tight time constraints. This problem becomes further complicated in the presence of sensor noise because noisy sensors provide, at best, incomplete information, which results in a distribution over target states to be intercepted. Since time is of the essence, to hit the target, the planner must begin directing the interceptor, in this case a robot arm, while still receiving information. We introduce an tree-like structure, which is grown using kinodynamic motion primitives in state-time space. This tree-like structure encodes reachability to multiple goals from a single origin, while enabling real-time value updates as the target belief evolves and seamless transitions between goals. We evaluate our framework on an interception task on a 6 DOF industrial arm (ABB IRB-1600) with an onboard stereo camera (ZED 2i). A robust Innovation-based Adaptive Estimation Adaptive Kalman Filter (RIAE-AKF) is used to track the target and perform belief updates.