Over the past decade, multirotor unmanned aircraft systems have seen tremendous growth in research and commercial applications. Much research has been done on developing complex control systems; however, due to current Federal Aviation Administration limitations, many systems can not be flown legally outdoors in the National Airspace. As an alternative approach to vehicle testing, this paper discusses the use of a six-axis load cell coupled with a physical quadrotor aircraft. The load cell measures the quadrotor-generated forces and moments, which are then integrated with the use of a forward dynamic model to provide real-time estimates of the system states. State information is telemetered from the simulation program back to the vehicle, overwriting onboard attitude/location estimates. To validate the use of a hardware-in-the-loop (HIL) simulation, load cell experiments were performed and compared with free-flight responses within an indoor motion capture facility. Step responses were performed in both configurations. Results show good agreement between the free-flight system and the rotational HIL simulation. Rotational response tests exhibited good agreement between simulated and experimental trials. Translational simulation step response data also demonstrated strong correlation with experimentally collected responses (both X and Z translations). In each of the tests, the added drag coefficient was relatively insensitive to the simulated response, providing additional confidence in HIL simulation as a path-planning testing tool with no/minimal experimental data collection.
展开▼
