Microbee is a four-rotor model helicopter based on the Mikrokopter system design. It's "brains" consists of 8 ARM-based microprocessors connected to various sensors and motors. It's equipped with the latest electronics hardware and is capable of photography, vision guided autonomous navigation as well as normal radio-control flying by a human being. An advanced vision recognition system is currently being developed based on the high performance, low power Texas Instruments OMAP3 platform.

All hardware & electronics is freely available in shops. The software is Open Source. Nothing is commercially-restricted.

Physical Characteristics

- Width: 40 cm (between opposite motors). 66 cm (propeller tip-to-tip).

- Height: 12 cm. (28 cm with HiLander).

- Weight: 700 grams (without battery).

- Flight Duration: ~15 minutes (with 167g/2200 mAh Lipo battery).

Electronics

- FlightCtrl ME v2.0.

- NaviCtrl v1.1 navigation controller.

- MK3Mag magnetic compass module for direction detection.

- MKGPS GPS receiver for navigation.

- MPX-4115A barometric pressure sensor for altitude estimation.

- Spektrum 2.4 GHz R/C receiver.

- Schulze alpha 5.35 35 MHz R/C receiver (modified with PPM output).

- Graupner/JR MX-12s 35 MHz R/C transmitter modified with 3 extra channels (1x 3-position switch, 2x trimpots) and Spektrum DM9 2.4 GHz module installed. Switchable between 35 MHz & 2.4 GHz.

- Radiotronix WI232EUR 868 MHz radio modem with 1/2 wave dipole & yagi antennas. Note that the 1/2 wave dipole antenna works best on-board as it doesn't require a ground-plane and has a reasonably omnidirectional radiation pattern. For long distance communication (up to 4km), a yagi antenna is used at the ground station.

- FB155BC Class-2 Bluetooth module.

- Maxbotix LV-EZ1 ultrasonic range finder.

- LIPO batteries 2200 mAh & 4400 mAh.

- An advanced computer vision system is being prototyped on a BeagleBoard single board Linux computer. It's a latest generation, low-power, embedded processing platform. Basic computer vision algorithms have been implemented on it. The source code is being optimised to off-load the number-crunching to the DSP & GPU processors to achieve real-time vision performance.

600 MHz ARM Cortex-A8 with NEON SIMD Co-processor
430 MHz TMS320C64x+ DSP (fixed point, six parallel vector units)
110 MHz SGX GPU

As the Beagle Board allows connection of keyboard, mouse, and display, it hosts the development toolchain and allows direct coding, compiling, and executing the embedded computer vision algorithms. Once stable, the smaller and lighter Gumstix Overo Water embedded module will be used on-board.