Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/C49D3B42-05F2-42EE-AED7-CBBBC8B453BF | ||
| Year | 2023 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Kleanthis Kyriazakis, "Design and implementation of an autonomous-ready electric motorcycle using Drive PX2", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2023 https://doi.org/10.26233/heallink.tuc.96669 | ||
| Appears in Collections |
The history of mass-production vehicles has been moving in two directions in recent years: electrification and autonomous driving. In fact, the combination of these two features looks set to dominate in the coming decades, as a large amount of resources is devoted to the development of these technologies. Responding to this trend, the motorcycle industry has recently been moving in the same directions, already presenting the first prototypes with impressive safety features in terms of balance and ride. The present diploma thesis proposes a low-cost, autonomous navigation system, designed for a prototype, autonomous-ready electric motorcycle that is being developed at the Technical University of Crete. The proposed system is based on Nvidia's Drive PX2 computing platform, which acts as the main brain of the vehicle, as well as two high-resolution cameras, which act as the main perception sensors. Using the Python programming language and the TensorFlow library, a real-time visual object detection system, based on a convolutional neural network, was developed, which specializes in vehicle and road traffic signal detection. At the same time, an Adaptive Cruise Control system was developed for the autonomous adjustment of the speed and the steering of the motorcycle, based on the perception of the current situation on the road. Information and data are displayed to the motorcycle rider via a touchscreen display, which offers various functions through a friendly graphical user interface, as well as analysis of specific scenarios. The interconnection of all subsystems, including the power supply from the motorcycle batteries, is achieved through a specially designed circuit and appropriate wiring. The proposed system is ready to be installed on-board in the prototype, autonomous-ready electric motorcycle, once its mechanical part is designed and tested for conventional driving.
Copyright © DIAS 2013
