Overview
In
realistic situations, planets, space and hazardous areas are dynamic, uncertain
and unpredictable. These unpredictable circumstances like adverse weather, exposure to radiation and toxic
wastes etc. limit
human’s capacity to carry out safe research and exploration
activities. This project focuses on providing a novel solution to safe
research and exploration in these environs through autonomous vehicle
technology.
The
sophistication of a modern and high-end vehicle requires more Electronic
Control Units (ECUs) interfaced with sensors and actuators to interact with the
physical world. In this project, an Autonomous Guided Vehicle (AGV) whose
multiple ECUs were networked using Controller Area Network (CAN) protocol, a
more robust and reliable in-vehicle communication protocol was built. The main
contribution of this study is the development and implementation of a fully
AGV with a reliable communication protocol and a surveillance unit to withstand
unforeseen circumstances in a dynamic environment.
This work also provides an additional research tool
to safe operation of automated vehicles in uncertain environmental conditions. It also makes useful contributions to wireless
ignition and autonomous path navigation for critical-safety systems in the
automotive industry. Findings from this project will stimulate further
research into AGV technology and possibly the development of more reliable automotive
protocols.
Profile
DEMO
Summary
This project sought to overcome the challenge posed by the
Free Ranging On Grid (FROG) vehicles by building an unguided vehicle. Autonomous
vehicles are the most suitable for carrying out tasks in dynamic, unstructured
and risky environments, eliminating the hazards humans are exposed to. Most available AGV are
low autonomy systems which require some driver interaction. This challenge was overcome by the fully
AGV developed in this project and therefore, did not require any map as a
guide. Thus, the
sensors and intelligent devices used in this vehicle are steps toward
developing a fully autonomous vehicle. A competent processor, such as the Mbed, that could perform
professional rapid prototyping was used to carry
out the complex intelligent tasks and the emergent coordinated activities of the vehicle.
The distributed control structure was implemented with
multiple Mbed processors networked to achieve an integrated solution of the vehicle’s
complex tasks. Communications between the six distributed ECUs were
realized on a CAN network. CAN is a highly reliable automotive protocol used in
safety-critical systems over a two-wire bus line and has a maximum bit rate of
1Mbit/s. The CAN bus reliability and fault tolerance features of this vehicle
were analysed by injecting faults into the vehicle ECUs. Being a robust
network, faulty node(s) disconnected from the bus, while others maintain reliable
communication.
The entire vehicle was powered by a 12 [V] rechargeable lead
acid battery which was regulated to 5 [V] required by each subsystem. Software
with an interactive GUI was created for the wireless start/stop controller
called BlueIgnition. The BlueIgnition controller was developed using a
ToothPick transceiver which has an inbuilt PIC microcontroller integrated with
a LinkMatik Bluetooth module. The controller was used to start or stop the
vehicle from a Windows PC via Bluetooth.
When the vehicle was started, it detected
and manoeuvred from any object that obstructed its exploration of the environment.
It also surveyed the environment using an on-board camera which captured images
of the obstructed areas and notified intrusion alerts
on the remote PC. The vehicle had the capability of acquiring and logging the
temperature, humidity and dew point of the explored environment onto the
PC using the Tera Term. The temperature condition of the
environment was used to control the vehicle’s speed. The vehicle’s activities
were diagnosed and communicated to the PC. When the exploration was completed, the vehicle was
remotely stopped and the captured images
stored
on the Mbed flash memory
were imported to a PC.
Finally, after some hardware debugging and software
restructuring, a fully autonomous explorer vehicle was achieved with a wireless
ignition system that presents a better
solution to the mechanical and remote control ignition challenges faced in the
automotive industry. A similar method was applied by Shaw and Politano in 2008 to
control home devices; however, this work creates a new automotive solution as there
was little to no literature indicting this application in vehicles.