Our thesis project focused on the development of a mobile robot, whose aim was the handling of industrial loads. In order to have high manoeuvrability, we chose a rhombic-like vehicle, using two motor wheels. The latter are controllable in steering and speed through independent electrical motors. In this manner, the mobile robot can easily perform lateral movements and move in tight spaces.
We were three of us working at this ambitious project that at the beginning was only on paper. We dealt with the mechanical project of the mobile robot, the choice of the sensors and the implementation of the various algorithms on the vehicle. We also developed a simulator in order to test the robot without moving it.
Mechanical structure
The mechanical structure of the robot consists of two aluminium plates supported by two motor wheels and two caster wheels. The two plates are joined by a steel shaft through four radial ball bearings. The aim of this solution is to allow the full contact between the ground and the motor wheels.
Logic scheme
The core of the system is the B&R APC 910, an industrial PC on which two operating systems (OS) are installed. The main one is ARWIN, a real time operating system developed by B&R. It performs the management of the CAN network, the vehicle control, the odometric localization and the sensor data fusion. The second OS is Windows 7 Embedded. The latter is fundamental in order to communicate with third-party devices and sensors, such as the laser scanner, the camera and the gyroscope. The communication between the two OS is performed through a virtual TCP/IP channel using the B&R PVI library.
The autonomous navigation ensures that the vehicle follows, in the best possible way, the predefined path. The control determines the values of the steering angles and the linear velocities starting from the output given by the localization systems. The sensors used for the localization are the encoders, which measure the steering angles and the rotation of the motor wheels, the gyroscope and the laser scanner, whose data are used to perform an Iterative Closest Point (ICP) algorithm. The interaction between the robot and the user is performed using a custom interface for planning and monitoring the task performed by the vehicle.
Our work
The prototype was realized ex-novo. We started from the project specifications, which required the capability to move a EUR standardized pallet (1200 x 800 mm), with a weigh of about 400 kg. The first step was the design of the 3D model together with the finite element analysis, in order to evaluate the dimension of the mechanical components.
The next step was the assembly of the mechanical structure and the electrical system.
After that, we started the key parts of our thesis: the development of the localization systems, the control of vehicle and the human machine interface.
After six months of work, we finished the development of the mobile robot and we started the test of the algorithms.
Acknowledgement
After this experience we would like to thank, first of all, Prof. De Cecco, who gave us the opportunity to work on this project. Moreover, our thanks go to Mattia and Alberto for the support that they gave us. Finally, we thanks the company Bellini F.lli of Rovereto for the hospitality during all the period of the development and the test of the robot.