| Idea Name: |
Autonomous Driving |
| Slogan: |
Safe Traveling for Everyone |
| Supervisor Name: |
Dr. M. Jameel Malik |
| Supervisor Designation: |
Assistant Professor |
| Supervisor School: |
SEECS |
| Supervisor Department: |
Electrical Engineering |
| Contact number: |
051-90852127 |
| Email ID: |
[email protected] |
| Abstract: |
| Millions of people losses their lives in car accidents. To have safe and sustainable driving fields, Autonomous driving is the only way forward. Our project focuses on making a cheaper and more effective prototype for autonomous driving.
The biggest hurdle in autonomous driving domain is its operation through limited geo fenced areas and we are planning to give cheap prototype for testing purpose |
| What is the unmet need in society that your idea will fulfill ? |
| Till today, the world has achieved maximum level four autonomy and the overall testing and deployment cost is really high. What our idea really focuses on is the small level cheap and portable prototype development of Autonomous driving according to road and traffic conditions of Pakistan. |
| Who needs it ? How many would benefit ? |
| The whole mankind needs it as Autonomous driving ensures safe traveling and cost efficiency for everyone. Millions of people losses their lives in car accidents and if we could develop efficient algorithms and safety protocols for autonomous operations, human lives could be saved. |
| How will the solution works |
| There are two parts of solutions: efficient algorithms and hardware deployment. For algorithm development, CNN is designed using NVIDEA'S model and whole algorithm is designed on Keras framework using OpenCv on Udacity simulator. The training and testing is also done on Udacity's simulator. For hardware deployment, Nvidea's Jetson Nano is used along with Racer Robots. For neural network training and testing, prototype paths are planned and segmentation techniques are used instead of Lane marking |
| Who are your competitors ? How is your solution different |
| Instead of making decisions on lane marking, we are using semantic segmentation techniques as in Pakistan, roads are not developed enough for self driving through lane marking. As it is evolving field, instead of performing testing on actual cars, experimentation could be performed on prototype. |
| Status: |
new |
| Entry Date & Time: |
2021-12-23 (1810) |
