LIDAR based autonomous car project

As the technology is growing and things are being automated day by day such as sensing alarms, locks, homes, and factories, so there is a need that our transportation system and vehicles should be automatic. The autonomous car or simply a driverless car is a robotic car which operates without the hu

Project Title

LIDAR based autonomous car project

Project Area of Specialization

Artificial Intelligence

Project Summary

As the technology is growing and things are being automated day by day such as sensing alarms, locks, homes, and factories, so there is a need that our transportation system and vehicles should be automatic. The autonomous car or simply a driverless car is a robotic car which operates without the human interaction. This type of car is capable of sensing the environment with LIDAR and navigating through that environment. We will also use NVIDIA Jetson Nano Xavier for image processing and cameras for the sign boards and traffic light detection. The major techniques to be used in our proposed project includes artificial intelligence, machine learning and computer vision. Also, sensor fusion and data fusion techniques are used in autonomous car to make it operate smoothly. The features / objectives of an autonomous car are the pavement lane keeping, obstacle avoidance, automatic braking and adaptive cruise control. This type of car is very useful for the society as it helps the disabled persons, children, aged-people or patients, etc. These type of cars will decrease the probability of accidents and traffic congestion and will increase the reliability of transportation.

Project Objectives

We want to make an autonomous car which will operate itself through AI and machine learning. There are four main functionalities / objectives of our car which are

1:Pavement Lane Keeping

To make a car that travels in between the given lane and follow the lane rules on different types of pavement.

2: Obstacles Avoidance

To make a car to avoid obstacles such that if there is some obstacle in its path it will take turns, or it can move towards the lane next to it. If there is no other lane and road has single lane, it will stop.

3: Automatic Braking

To make an automating braking system of a car which applies breaks when there is an obstacle in front of it. The obstacle could be a human, animal or any other vehicle etc.

4: Adaptive Cruise Control

To achieve adaptive cruise control in our car. Adaptive cruise control means that the car itself decides the speed, can control the speed by itself and can break itself without the engagement of any person / driver.

 We want to achieve the level 3 of autonomy in which most of the work is done by the machine, but driver is necessary to take over in complex and uncertain situations where machine gives control to the driver. With all these objectives / functionalities, driving capability of the driver such that turning, cruise controlling, and fuel consumption has been enhanced and improved.

Project Implementation Method

As we are going to implement the prototype of the autonomous vehicle and  for this purpose,

• We will use a servo motor based remote control car with the LIPO battery power.
• We will use donkey car library of python which is an open source self-driving platform for small scale cars. Donkey-distribution is a high-level self-driving library. We will install the Donkey car software on the robot platform (ROS) and we will implement it on NVIDIA Jetson Nano Xavier.
• Next we will train the remote control car to drive itself. It uses a supervised learning technique often referred as behavioural cloning  and we will use deep learning for image classification and object detection.
• Further we will use a camera to detect the obstacles coming towards the car, detecting the sign boards, pavement lanes, traffic lights, pedestrians and for avoiding some common types of obstacles faced in daily life.
• For turns, we will use the adaptive cruise control for avoidance from any type of car accidents and for maintaining the speed of car during running on the road in different situations.
• We will use other sensors like RP-LiDAR A3 in which the core of the sensor runs clockwise to perform a 360-degree omnidirectional laser range scanning for its surroundings and then generate highly accurate geographical data in a 3D map. LiDAR generate the point cloud of the environment around the sensor. This point cloud gives different 3d shapes. So, we will use these points to make the classification of objects. Laser beams are used in lidar that hit to the obstacles and come back with the speed of light. We calculate that time and get the exact distance from the car to that obstacle.
• Ultrasonic sensor is used to support the car in parking as collision with other cars and nearby obstacle detection like pits and speed breakers. It works by sending out a sound wave at a frequency above the range of human hearing so, by calculating the travel time and the speed of sound, we will calculate the distance.
• Finally, by using different data fusion and sensor fusion techniques we will try to develop more consistent and accurate autonomous car.

Benefits of the Project

There are many benefits of this project socially, technologically, economically and environmentally. The social benefits could be that the people with disabilities like handicapped and blind persons can use this to go from one place to the other. The people who don’t drive or don’t know how to drive can get benefit from this. It could be beneficial for children and old aged people for travelling. The technological benefit is that, this project shows technological advancement in transportation and power of artificial intelligence and machine learning. Economic benefits are that it saves money spent on traffic regulations and to control the air pollution and it also saves the fuel up to 10% as studied by Ohio University. Autonomous cars have a positive effect on the environment also such as controlling the air pollution, a smaller number of accidents, dropping the emissions by 60% which are produced by traffic congestion. Autonomous cars also helpful in maintaining the time as it reduces the travelling time up to 40% as according to KPMG’s Connected and Autonomous Vehicles – The UK Economic Opportunity. If all vehicles on the roads are autonomous, they become interconnected and it will save them from collision and accidents. According to the study, up to 90% of road accidents occur due to the negligence of drivers and through this technology, we can reduce it and can save many human lives. Another benefit could be that the people will start to obey traffic rules and regulations as the vehicles are handled and controlled by the machine and it will strictly and most precisely follow the instructions given to it.

Technical Details of Final Deliverable

Our final product will be unmanned car with some major functionalities of high automation (level 5 of autonomy). Our product will be based on NVIDIA Jetson nano Xavier which will operate the car using robot operating system, sensor fusion and data fusion. Robot car will navigate through the pavement by using deep learning, effective detection and identification of moving obstacles, sign boards and traffic lights by taking input using the camera, using machine learning in backend, road mapping for safe driving using LiDAR and ultrasonic sensors for avoidance of collision with other cars, avoidance from pits and handling the speed breakers. We use adaptive cruise control technique for controlling the speed of car in certain situations and applying breaks in uncertain situations. So, for safety the speed of our car will be limited to 25km per hour.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

IT

Other Industries

Transportation

Core Technology

Artificial Intelligence(AI)

Other Technologies

Robotics

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com