Self Driving Car Using Visual EgoMotion Estimation

Self-Driving Cars are currently manufactured by Ford, Tesla and General Motors. They are investing millions of dollars in autonomous vehicle driving research. The benefits of such a car is fully autonomous removing the need of a human driver. According to recent reports of accidents from all

Project Title

Self Driving Car Using Visual EgoMotion Estimation

Project Area of Specialization

Robotics

Project Summary

Self-Driving Cars are currently manufactured by Ford, Tesla and General Motors. They are investing millions of dollars in autonomous vehicle driving research. The benefits of such a car is fully autonomous removing the need of a human driver.

According to recent reports of accidents from all over the world it has been seen that every year millions of people died in road accidents. And in study it has been observed that 87% crashes are due to driver factors. So, in order to reduce these crashes and save lives a number of companies are creating these kinds of cars but the cost is very high. Two years ago, google has managed to create the world’s first autonomous car. The Google car's problem is it uses a very expensive ($7500) Radar.

A middle-class human/family can’t afford it The Radar is used to recognize the environment and create a High-resolution map. Our solution is to use cameras instead of Lidar and radar. And implement visual Ego-Motion Estimation algorithm to localize the car.

The purpose of this project is to create a fully autonomous car which should be able to reduce the road traffic injuries by using visual ego-motion estimation algorithm with 4 cameras and minimal overlaping field of view. Additionally, a driverless car can reduce the number of traffic jams, avoiding human errors, allowing disabled people to drive long distances.

Project Objectives

The objective of this project is to design a self-driving car employing visual ego-motion estimation algorithm that rely only on cameras instead of devices like radar and Lidar. Radar and Lidar are used currently by Google self-driving car.

In short, the main goals of the projects are:

• Develop an ego-motion estimation algorithm using generalized multi cameras system.
• Develop the algorithm to operate the basic functions of a self-driving car.
• An Investigation and a practical solution for the general cases of intra camera correspondences.

Project Implementation Method

The main problem to be addressed in this project is to develop a self-localization and mapping algorithm. As described earlier many car companies uses sensors like RADAR and LIDAR to localize the surroundings. Although these cars also use cameras but these cameras are just to assist the driver and are used during parking. As a solution to this problem in this project 4 cameras will be used and visual ego motion estimation will be implemented for minimal overlapping camera system, localization will be done using data provided by the cameras. A prototype car will be equipped with 4 cameras and minimal overlapping field of view.

Benefits of the Project

• Self-driving vehicles can help reduce driver error.
• Higher levels of autonomy have the potential to reduce risky and dangerous driving behaviors.
• Highly automated vehicles can help people with disabilities, like the blind to live the life they want.
• These vehicles can also enhance independence for old people.
• Ride-sharing could reduce costs of personal transportation, providing more affordable mobility.
• These cars can help to avoid the costs of crashes, including medical bills, lost work time and vehicle repair.
• Fewer crashes may reduce the costs of insurance.
• In a fully automated vehicle, all occupants could safely pursue more productive activities, like responding to email.

Technical Details of Final Deliverable

A prototype car will be equipped with 4 cameras mounted at the front, the rare, the right, and the left sides with minimal overlapping field-of-views. Other Technical details are as follows:

• CPU: Intel Core I3 with peripherals.
• Position controller with shaft encoder.
• Speed controller.
• 4-Cameras: 12 MP

Dimensions of the project are as follows:

• Weight:  15 Kg
• Length: 3 ft
• Width: 1.5 ft
• Hight: 1.5 ft

Final Deliverable of the Project

HW/SW integrated system

Type of Industry

Manufacturing , Transportation

Technologies

Artificial Intelligence(AI), Robotics

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

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