Autonomous Vehicle

Path Planning in C++ Part 2

I managed to solve the driving-in-a straight line by calculating the trajectory of the vehicle instead of directly feeding the trajectory as I had done earler. This can be seen in the image below. But now I have developed another problem such that my vehicle is hitting other cars on the same lane. I thought I would be able to solve this by reducing the value of dist_inc down to 0. … »

Path Planning on C++

Introduction The path planning project is by far the most interesting project after Behavioral Clonning Project. This project is for Term3 (Last Term) and it is the first project for this term. The main objective is to create a path planner that will be able to safely and smoothly create a path for a vehicle to follow along on a 3 lane highway traffic.It should be able to keep inside its lane, avoid hitting other cars and pass slow moving traffic all by using localisation, sensor fusion and map data. … »

Data Collection Monday

I go to BSF on Monday evenings to Westlands St Marks Church and the drive there with my friends provides a good chance to collect some data for my self driving car programme. My data collection device is still my mobile phone mounted on the dashboard. I have explained how to do this in my previous post. Yesterday’s run was like normal but over the last 3 weeks, my videos have been appearing with some darkness which is as a result from the mounting angle facing the road. … »

Collecting Driving Data On Kenyan Roads

Ever since I dipped my mind into the autonomous vehicles world, I have been noting carefully road transport in a whole new dimension. I now see data and need lots of it. Training a CNN needs alot of data and even when I was working on the Lane Detection Algo project, I used data that was provided by (Udacity)[]. But now I deceided to put the same algo on local roads here in Kenya. … »