Visual Navigation Robot
OpenCV on the Raspberry Pi detects human faces. I modified the script to move the robot left or right based on the coordinates of the detected face so that the face remains in the center of the camera's view. The robot will move toward the detected face until it is a certain distance away then stop. This is the earliest working prototype of this project, much more to come soon.
I'm using a Ravpower 5200mAh USB battery pack to power the Raspberry Pi and a 6xAA battery holder that came with the Pololu RP5 tracked robot chassis to power the motors.
Motor controller is the SN754410NE powered from the Raspberry Pi's 5V GPIO pin and using 4 GPIO pins to control the chip. The first attempt was using Quick2wire's Raspberry Pi GPIO admin/python software and then moved to pi-blaster (PWM output on all GPIO's) because I will be adding a servo or two in the future.
Edimax EW-7811Un USB WiFi adapter to SSH into the robot.
Update 7-9-13: Running the python script at Raspi boot by adding it to /etc/rc.local.
Update 8-9-13: I'm using this robot as a platform for various visual navigation code using OpenCV so please note the first video here uses the face cascade and object recognition. The second video uses real time point detection, followed by grouping based on distances then navigation in the direction of the area of interest. IR sensors are used in the second video to stop the robot when close to the detected area. The third video explains the details of the hardware itself.