Control of Systems with MEMS Sensors and

 Actuators via Data Mining Techniques
Project info

Keywords

Project Summary

Delta Wing Background

Wind Tunnel Experiments
 -Delta Wing Aircraft Model
 -MEMS Bubble Actuator

Mining Actuation Patterns

Conclusion
Project Award Information
  • NSF Award Number: IDM- 0097438
  • PI: Wesley W. Chu, Computer Science Department, UCLA
  • Co-PI: Chih-Ming Ho, Department of Mechanical and Aerospace Engineering, UCLA        

Keywords

  • MEMS sensors and actuators

  • Dynamic control

  • Delta wing flight control

  • Temporal and spatial data mining

Project Summary

     This is an interdisciplinary project between the Computer Science and Mechanical Engineering Departments. Processing and interpreting the vast amounts of sensor information in determining the actuation schema to control a system is an open problem. To address this problem, we use a novel data mining technique to derive classification rules and association rules from training datasets that consist of multivariate input and output variables. Based on the system-operating environment, the best applicable rule can be selected to derive the actuation schema that drives the system to the desired state. The input-output relationship for the delta wing aircraft is highly non-linear, and the transfer function between the sensors and actuators is especially complicated. We plan to use  delta wing aircraft input/output MEMS test bed samples to develop a scalable data mining technique that discovers full input-output relationships under a wide range of conditions (dynamic, temporal, spatial, etc).

     This project leverages on our past data mining research of multivariate variable training datasets, as well as the available MEMS sensors and actuators for UAV (Unmanned Aerial Vehicle) application and wind tunnel measurement facility to collect the training data. Based on the collected experimental data for dynamic system behavior, our current research extends the mining algorithm for summarizing the relationship between actuation and flight dynamics, and developing the rule selection strategy for actuation schema so that the aircraft can be navigated to follow a specified trajectory. These results provide insight into the behavior of flight dynamics of the delta wing, actuation, and operating environment.

Delta Wing Background

  • nA sweep of wind produces pairs of vortices above the Delta wings.
  • nThese vortices are sources of low-pressure flows that provide “suction” which produces a portion of the lift for the aircraft.
  • nBy controlling the location of these high “suction” vortices, the lift and moment about the aircraft can be controlled.
  • aActuators control the “vortex” to provide forces and moments for flight control of the delta wing aircraft.
  • Based on the operating environment such as wind speed and angle of attack of wind to the aircraft, actuators are used to control the flight trajectory.n
  • nActuator configuration (Actuator # and position)  is used for flight control (input). Sensors are used to monitor the corresponding state (e.g. forces and moments) of the aircraft (output).
  • nData mining techniques are used to summarize the nonlinear input/output relationship between actuator control and the sensor output data sequences.

Wind Tunnel Experiments

nThe data is collected from wind tunnel experiments. We first design the devices nfor the experiments:
n

Delta wing aircraft model:

nAutomatic movement mechanisms simulating the transient conditions for the delta wing aircraft:
  • nMotion control hardware
  • nActuation mechanisms

nPerform experiments to collect static and dynamic data for the flight dynamics (i.e. forces, moments, etc.) with different actuation positions.

nMEMS bubble actuator:

Linear MEMS bubble actuator arrays installed on the leading edge of the delta wing model (shown inflated). 

 

 

Mining Actuation Patterns

 n     

n     For a given actuation sequence, a flight trajectory can be generated. nDifferent actuation sequences can generate different flight trajectories.nEach point of a given trajectory corresponds to an actuation based on its precedent’s sensor output. nEvery actuation decision is based on operating environment information, such as  angle of attack of wind to the aircraft, velocity, position (x,y,z), etc. nA family of trajectories can be generated by different actuation sequences.

     For a given operating environment, by selecting the desired trajectory, we can obtain its corresponding actuation sequence. nThe trajectory shown in the figure corresponds to a sequence with five actuations <2_90, 1_90, 2_95, 1_120, 2_110>, where 2_90 is an actuator configuration (actuator #2 actuated at 90 degree position).

n     Similar trajectory path segments can be classified into groups and their corresponding actuation sequences can be extracted. nPatterns can be derived from the actuation sequence to show the effect of actuation (such as lift, drag, rotation) on flight trajectories.

 

Conclusion

  • Derived the inter-relationship of the behavior of flight dynamics of the delta wing aircraft with actuation sequences and operating environment.

  • Data mining technique is feasible to summarize the complex and nonlinear relationship of actuation sequences and operating environment with flight trajectory.

  • Such relationship provides insight of flight control for delta wing aircraft.