This paper presents the autonomous micro aerial vehicle pilot, a new autopilot platform weighing 6.25 g and measuring 11.3 cm 2 , specifically designed for use on micro/miniature aerial vehicle mobile sensing platforms. An overview of the hardware, firmware, ground station, and validation testing used to demonstrate this autopilot as a viable research instrument for atmospheric thermodynamic sensing on micro/miniature aerial vehicles is presented. The autonomous micro aerial vehicle pilot incorporates a 16 bit 140 MHz processor, global positioning system, dual radios, inertial measurement unit, pressure sensor, humidity sensor, and temperature sensor. Through these components, the autonomous micro aerial vehicle pilot is capable of full-state feedback, vehicle state estimation, localization, and wireless networking. Notable features of this autopilot are its dual-radio configuration, providing redundancy and adaptability in communication, and the detachable sensor breakout design that allows for increased flexibility in sensor placement on a vehicle. Full-state feedback of the autopilot platform was validated through a series of bench tests. This includes a unique technique for dynamic inertial measurement unit validation performed using the present group's model positioning system as well as comparing estimated pressure values with known values at multiple heights and global positioning system values with a known path. Systemwide validation was performed through flight tests on a micro/miniature aerial vehicle airframe. I. Introduction M ICRO/MINIATURE aerial vehicles (MAVs) are becoming increasingly popular in civil, military, and scientific fields. MAVs can be used for a variety of purposes, such as navigating compact spaces like those found in urban environments, search and rescue, and disaster assessment. As a result, research communities are constantly experimenting and pushing the bounds of MAV capabilities. Our research group is an example of such, having worked extensively with MAVs to develop cooperative control [1,2], multihop communications and sensor networking [3,4], dynamic data-driven application systems [5], and extreme weather monitoring applications [4]. Our group's previous autopilot, the Colorado PIC (CUPIC) [6], developed at the University of Colorado at Boulder and shown in Fig. 1, served as the core platform of our autonomous research activities until 2014. The CUPIC was designed, based on the capabilities available a decade ago, to be a simple autopilot equipped with a limited number of sensors while still allowing system observability for a MAV under severe weight, size, processing, and sensor capability restrictions [7]. In our group, the CUPIC was predominately flown on a 0.94-m span delta wing aircraft (shown in Fig. 2) and used to test swarm control [1] and sensor networking [3]. During the seven years since the CUPIC's last update, a number of technological advances have been made in the areas of embedded systems and microelectromechanical system (MEMS) sensors. Accordingly, we deci...

No abstract