Abstract: Research has shown that erratic human behaviour such as impaired driving, drugged driving, unbelted vehicle occupants, speeding and distraction are factors in as much as 94% of the crashes in roads today. Automation in this field has the potential to enormously reduce the incidence of such crashes. Higher levels of autonomy mean that the problem of driving is no longer one that human drivers have to solve. This is an area that has gained considerable traction in recent years. The modus operandi in the research community is to understand human driving behaviour and build autonomous units that can imitate this behaviour. This complicated issue domain calls for elaborate solutions that frequently involve numerous modules operating in unison. Each of these modules deals with a particular issue and transmits its solutions to the succeeding modules for processing. The vehicle's controller component, which carries out the predetermined behaviour, receives the ultimate outcome. Additionally, since everything must occur in real-time, prediction speed is just as crucial as underlying accuracy.
This sophisticated modular design has been discovered to be inefficient, and deep learning has been found to be a good replacement. Deep learning involves automatically learning complex mathematical functions that characterise a specific domain. Understanding human drivers is a complex task. It involves emotions rather than logic, and these are all fuelled with reactions. It becomes very uncertain what the next action will be of the drivers or pedestrians nearby, so a system that can predict the actions of other road users can be very important for road safety. The car can observe, gather all the information it requires, and interpret it thanks to a 360-degree vision of its surroundings. Once the data is loaded into the learning system, it can think of every possible move those other drivers could make. It resembles a game in which the player must choose the best move from a limited number of options in order to beat the opposition. In this issue area, an autonomous unit's functions include localising the vehicle in its environment, enhancing perception, and actuating kinematic motions in self-driving cars. This guarantees both easy commuting and road safety.
|
DOI:
10.17148/IJARCCE.2022.11789
