Stair Climbing via Successive Perching

Abstract

Stairs are a primary challenge for mobile robots navigating indoor human environments. Stair climbing is a useful, if not necessary, capability for mobile robots in urban search and rescue, security, cleaning, telepresence, elder care, and other applications. Existing stair climbing robots are large, expensive, and not always reliable, especially when descending stairs. In this paper, we present a novel approach for stair climbing that is achievable by a small mobile robot with minimal actuators and sensors and, thus, cost. The proposed robot has articulated tread assemblies on either side of a chassis. Using feedback control, the robot can balance on the edge of a single step. As the robot drives up the step, the chassis pivots to maintain the center of mass directly above the contact point. The dynamics of the system are derived with the Lagrangian method, and a discrete-time integral controller with friction compensation is designed to stabilize a stair climbing trajectory. The algorithms used to estimate the state of the system with low-cost noisy proprioceptive sensors are explained in detail. No external motion capture system is used. Simulation results are compared with successful experimental results.