Ignore Drift, Embrace Simplicity: Constrained Nonlinear Control through Driftless Approximation
By: Ram Padmanabhan, Melkior Ornik
Potential Business Impact:
Guides robots to reach goals with simple steps.
We present a novel technique to drive a nonlinear system to reach a target state under input constraints. The proposed controller consists only of piecewise constant inputs, generated from a simple linear driftless approximation to the original nonlinear system. First, we construct this approximation using only the effect of the control input at the initial state. Next, we partition the time horizon into successively shorter intervals and show that optimal controllers for the linear driftless system result in a bounded error from a specified target state in the nonlinear system. We also derive conditions under which the input constraint is guaranteed to be satisfied. On applying the optimal control inputs, we show that the error monotonically converges to zero as the intervals become successively shorter, thus achieving arbitrary closeness to the target state with time. Using simulation examples on classical nonlinear systems, we illustrate how the presented technique is used to reach a target state while still satisfying input constraints. In particular, we show that our method completes the task even when assumptions of the underlying theory are violated or when classical linearization-based methods may fail.
Similar Papers
Exploiting Over-Approximation Errors as Preview Information for Nonlinear Control
Optimization and Control
Makes robots smarter by using mistakes as clues.
Reachable Predictive Control: A Novel Control Algorithm for Nonlinear Systems with Unknown Dynamics and its Practical Applications
Robotics
Teaches robots to move without knowing how they work.
A constrained optimization approach to nonlinear system identification through simulation error minimization
Optimization and Control
Makes computer models learn faster and better.