Manoeuvre Recognition Using A Low-Cost Standalone MEMS-IMU System

This paper describes the use of a low-cost standalone MEMS-IMU (micro-electromechanical system - inertial measurement unit) sensor system developed by the National Research Council Canada (NRC) for manoeuvre recognition in helicopters. The system records accelerations, angular rotation rates, magnetic flux, altitude, location and velocity through its IMU and GPS. The MEMS-IMU system was flown on the Bell 206 helicopter operated by the NRC Flight Research Laboratory in two unscripted flight tests. Comparison of the system's measurements with those of the helicopter's inertial navigation system (INS) shows very good agreement, demonstrating that the system functions well in the intended environment and its measurements provide accurate and meaningful data. Data visualization of the flight test data mapped into a 3D virtual reality representation is presented to provide insight into the structure of the flight data. Manoeuvre recognition based on the sensor system's measurements was attempted using a data-driven classifier model. Even with different subsets of the MEMS-IMU measurements, the classifier results show that the parameters recorded by the system are sufficient for accurate manoeuvre classification.