Print

Carpenter, Gregory P.

Electrical Engineering

2012

MS

In recent years advances in low-power embedded computing and sensing hardware have led to the ubiquity of two increasingly important technologies: wireless sensor networks (WSNs) and Micro-Aerial Vehicles (MAVs). Improvements in size and energy efficiency have allowed for the deployment of WSNs in increasingly exotic and remote environments for a range of sensing missions from monitoring environmental parameters to structural health in civil structures. WSNs however still suffer two substantial limitations: (1) available energy from on-board power resources, which constrains broadcast radius thereby limiting the distance collected data can be transferred, and (2) fixed location which limits the spatial resolution of the data that can be collected by each sensor node. Given these limitations, the spatial range of MAVs and their ability to replenish their energy supply make MAVs an ideal complement to WSNs, both as a data 'mule' increasing the radius which WSN data can be transmitted, and as a mobile sensor node able to collect data over a wide range of spatial positions.
This work has explored three specific building blocks of a MAV/WSN coupled 'system.
(1) An analytical study of the strategy of error-constrained frequency selection for WSN collaborative beamforming (CB). This method used bounds on error in location estimation of individual sensor nodes, and bounds on synchronization error between nodes to select an operational CB frequency that achieves desired effective gain characteristics. This selection of CB frequency increases the effective range and directionality of the transmitted signal that a WSN can use to establish a communication link with an MAV. (2) Using the Sentilla Tmote Sky wireless sensor, a hardware implementation of WSN consensus forming was developed to allow for data aggregation and fast estimation of shared WSN parameters. This consensus forming technique was further extended to the task of spatial centroid estimation of the WSN, a necessary parameter for CB.
(3) Finally, using a Gumstix Overo embedded Linux operating system (OS) computer along with an on-board Microstrain 3DM-GX325 inertial measurement unit (IMU), a hardware/software module as the 'brain' platform for the MAV was developed. This 'brain' module was designed to record inertial state information of the MAV using the IMU, while concurrently recording (in one mode of operation) or generating (in another mode of operation) the standardized pulsed-position modulation (PPM) control signalling widely used in many classes ofMAV. The specific MAV used in this work is the Mikrokopter Vertical Take-Off and Landing (VTOL) quadrotor helicopter.