Here’s a preliminary demo of some LIDAR highway vehicle/pedestrian detection algorithms I’ve been working on the past few months. This method is an algorithmic approach compared to a training/learning based method, e.g. neural networks. I think there is some interesting discussion on the performance differences between the two approaches, and hope to publish about this soon.

I’ll also be working to develop vehicle classification and tracking algorithms so that I can provide predictions of where the vehicle may be at some timesteps ahead.

Here’s a quick demo of Hybrid A* in CARLA. Currently, it is using the velodyne_height_map plugin to detect obstacles, and I create a costmap/occupancy map based on that. For the Hybrid A Star implementation, I chose one done by @karlkruzer due to its easy integration with ROS. I’ll post a GitHub / full write up later once I get control implemented.

Learning logistic regression can be confusing the first time around. One of my issues early on was working on through the different notations you could have based on the classifiers you chose, e.g. (plus/minus one) vs. (zero/one). This won’t be a post explaining logistic regression (you can find tons of Medium articles on that), but rather the two different ways you can express your cross-entropy loss function.