Abstract
This paper proposes a low-complexity, fast sensorless collision detection method for permanent-magnet synchronous motors (PMSMs) based on a forgetting-factor recursive least-squares (FFRLS) estimator. The method estimates the load torque from the motor motion equations and detects collisions by monitoring abrupt changes in a processed load-torque metric. The algorithm is tailored for resource-constrained embedded controllers: it avoids high-order matrix operations and the use of external sensors while achieving rapid detection. Experimental results under constant speed, acceleration, and time-varying load conditions demonstrate fast and repeatable detection performance. The approach provides a practical trade-off between computational cost and detection reliability for embedded motor drives.