Hardware
“AutoNOMOS Mini” is a model vehicle (scale 1:10) developed for educational purposes. The car can be controlled using a cellular phone or can be programmed to drive in fully autonomous mode. The main computer is an Odroid board (XU4 64GB) running Linux and the Robotic Operating System (ROS) on top. The chassis of the vehicle comes from the model-car building community. The vehicle has been motorized with a 15-Volt Faulhaber brushless motor with its own PID controller. Steering is done through a servo motor XciteRc XLS-19s. See Fig. 1 for the details of the assembled unit.
Figure 1: View of the lower layer of components
Electronics and sensors are arranged in two layers. Fig 1 shows the lower layer, Fig. 2 the assembled car. The sensors in the vehicle are a rotating laser scanner (RPLidar 360) which provides detection of obstacles around the vehicle. A Kinect-type stereoscopic system, as well as a fish-eye video camera pointed to the ceiling, have been mounted on top of the car’s body. The Kinect sensor provides a cloud of 3D points that can be used to detect obstacles. The video camera in the Kinect-type sensor can be used to detect the lane and objects in front of the car. The camera pointed to the ceiling can be used to identify markers, providing in this way a simulation of a GPS navigation unit, so that the car can localize indoors. The sensors are all connected through the USB ports. A 4-port US-Hub provides additional USB connectivity for the Odroid board.
Figure 2: View of the complete assembly (with the Kinect-type sensor and the wide-angle videocamera)
The car has LEDs for simulating the head- and tail-lights, as well as blinkers. Fig. 3 shows the main connections between the different electronic modules. The Odroid board is the main computer. It connects through USB to the Kinect-type sensor (Intel camera), the Fisheye camera and the WiFi module (through a USB Hub). The Faulhaber brushless motor and the RPLidar connect through a serial interface. The FT4232H module is a USB to UART/MPSSE board -- it converts a USB port into four serial ports. One of the serial ports connects to the RPLidar, a second one to the Faulhaber controller, and the third to an Arduino Nano board. The FT4232H module is therefore just a converter for interconnecting the mentioned components to a USB port.
The Arduino Nano is used to switch the vehicle on and off (using its digital pins and relays), to control the LEDs, and to monitor the battery (through the ADC pins). The Arduino board sets the steering angle using a PWM signal sent to the steering servo motor. An additional board, the MPU6050, provides measurements from accelerometers and gyroscopes, that can be used to complement the odometry and also to measure the vehicle’s rotation. The MPU6050 is connected to the Arduino board through the I2C pin.
Figure 3: Electronic Modules and interconnections
Figure 4: The tree of interconnections
If you push the green button on the car (Fig. 5) for 3 seconds, the Arduino board will enable one of the relays in the car providing thus 5 Volts to the main board. The status light (the light between the red and green buttons) will become yellow. All components (Odroid, servo, Lidar, lights, etc.), except the brushless motor, will be enabled. The Odroid starts booting. If you push the green button again, it will enable another relay that provides 15 Volts to the brushless motor. The status light will then become green. By pushing the green button, you can toggle the 15 Volts on and off, in order to connect or disconnect the brushless motor from the battery. Pushing the red button turns off the main board and sensors. The status light will become red.
Figure 5: Green button: switches on the 5V and 15V circuits. Red button: stop, it shuts down the car. Status light: red: 0 V, yellow: 5 V, green: 15-13 V).