STTs for Temporal Reach Avoid Stay (T-RAS) Tasks

A sampling-based STT framework with closed-form control for unknown nonlinear systems

Temporal Reach–Avoid–Stay (T-RAS) tasks appear across robotics and safety-critical autonomous systems: reach a target set within a desired time while avoiding unsafe regions and remaining in a safe state space. Classical tools such as Hamilton–Jacobi reachability or control barrier functions struggle with scalability, model uncertainty, or online computation costs.

This project develops a sampling-based Spatiotemporal Tube (STT) framework with an approximation-free, closed-form controller that allows unknown nonlinear systems to satisfy general T-RAS specifications.

The STT generation framework.

Omnidirectional Robot — Safe Navigation

The robot safely reaches the target within the prescribed time while avoiding all obstacles.

Magnetic Levitator — Time-Varying Unsafe Sets

Handles time-varying unsafe intervals while reaching the target at the specified time.

Related Publication

Spatiotemporal Tubes for Temporal Reach–Avoid–Stay Tasks in Unknown Systems
R. Das, A. Basu, P. Jagtap
IEEE Transactions on Automatic Control (TAC), 2025

Read the paper

References