Tuesday, August 27, 2013

Quanser 3 DOF Helicopter Platform Helps Develop High-Speed Embedded MPCs

Were you to search the Internet or YouTube for videos on applications of specific Quanser devices (a practice we definitely recommend), you’d find they’re in wide use by researchers and educators all around the world. 

One of the videos we’ve seen lately involves our 3DOF Helicopter.  It’s being used by Jonathan Currie and his team at the Industrial Information and Control Centre housed at Auckland University of Technology (AUT) in New Zealand.  

They’re involved in developing a high-speed embedded Model Predictive Controller (MPC), a constrained optimal control strategy which has been widely used in the world of chemical and process control. Like other researchers, they believe that MPCs have considerable applicability within the high speed (kHz) embedded world of unmanned vehicles (air, ground and sea).

Jonathan and his team chose to test their embedded MPC control with the Quanser 3 DOF Helicopter, because they were seeking a challenging, nonlinear MIMO plant for which traditional control strategies (such as PID) were either unsuitable, very hard to tune or hard to get going at high speed. They implemented their MPC algorithm on a low-cost Texas Instruments Delfino microcontroller together with an auto-coding toolset implemented in MATLAB®.

Quanser’s 3 DOF Helicopter provided a robust, open-architecture, and visually impressive platform. The research team could easily remove the Quanser Q8-USB DAQ from the original solution (where the controller runs on the PC) and, when coupled with the Quanser VoltPAQ-X2 power amplifier, measure and control the system using the TI microcontroller. The 3 DOF Helicopter enabled them to test their own controller algorithms and hardware on real system and validate their control design.

Currently the AUT team is building, generating, compiling and deploying MPC controllers capable of 5 kHz sampling rates in as little as 10 seconds on microcontrollers that only cost a few dollars. This proved so promising that the team has started a spin-off company, Inverse Problem Ltd, to commercialize the ideas. Quanser is pleased that its products contributed to the AUT team’s research.

No comments: