UNLV Autonomous Vehicle: Our dedicated team has been deeply involved in advancing autonomous vehicle technology. I have had the privilege of leading a group of highly motivated graduate and undergraduate students in this endeavor. After over a year of dedicated effort, we are now witnessing promising results. We have successfully implemented self-driving technology on our university campus, utilizing a customized Lincoln MKZ equipped with an array of sensors, including 4 LiDARS, 5 RADARS, SONAR, 4 Lucid vision cameras, a thermal camera, IMU, GPS, and drive-by-wire technology. As for our software stack, we've tailored open-source Autoware Universe to accommodate our sensor suite, enabling the vehicle to operate autonomously within its Operational Design Domain (ODD). Our self-driving car rigorously adheres to all traffic regulations, including stop signs, traffic lights, right-of-way, parking, and more. Learn more.
HD Map for autonomous driving: High-definition (HD) maps play a pivotal role in enabling autonomous cars to navigate and operate safely. These intricate maps provide detailed and precise information about roadways, traffic signs, lane markings, intersections, and other critical features. By integrating real-time data from various sensors with static information from HD maps, self-driving vehicles can make informed decisions, plan routes, and navigate complex environments with enhanced accuracy and reliability. HD maps are a fundamental component in the quest for safer and more efficient autonomous driving solutions. Learn more.
Sensor Calibration: Sensor calibration is an integral part of any robotics project. Through calibration, we get the extrinsic parameters of those sensors with respect to each other, for example, the XYZ (translation) and roll, pitch, and yaw (orientation) of one sensor with respect to another one. Thus, we can transform data from one coordinate system to another one like a detected object is 10m away from the lidar and 8m from the car coordinate frame. Learn more.
Autonomous car blind spot: Our lab team worked to solve an autonomous car blind spot issue. A self-driving car depends on its sensors to perceive the surrounding environment. Those sensors have their own advantages and drawbacks. For example, if there is a wall beside the road and a pedestrian suddenly comes from behind the wall with the intention to cross the road, the car can not see him. Thus, to avoid potential accidents, we developed a system that can detect that hidden pedestrian and notify the car wirelessly. Learn more.
Bosch Future Mobility Challenge 2022: Our team at UNLV participated in BFMC 2022. We were given a 1/10 car with a raspberry pi 4 processor and PiCamera. The car should navigate through a smart miniature city autonomously. We employed traffic sign detection systems as well as lane detection systems. We set up the track in our lab room to test the car in real-time. Our car can successfully drive itself on that track and we got selected for the final event as a result. Learn more
Wrong-way vehicle detection from traffic footage: Wrong-way driving is a significant contributor to road accidents and traffic congestion worldwide. By effectively identifying wrong-way vehicles, we can greatly reduce the number of accidents and alleviate traffic congestion. In this project, we present an automated system for detecting wrong-way vehicles using surveillance camera footage captured on the road. Our system operates in three stages: vehicle detection, tracking, and wrong-way vehicle identification. Learn more
Skin Lesion Classification: Skin cancer is recognized as the most common kind of cancer in the world. It could be deadly if not identified at the primary stage, which makes early detection very crucial. In this project, we present a novel approach for classifying seven types of skin lesions using an ensemble learning-based model with weighted averaging. The ensemble is constructed using five deep neural network models, namely ResNeXt, SeResNeXt, ResNet, Xception, and DenseNet. Our models are trained and evaluated on a dataset of 18,730 dermoscopy images from the HAM10000 and ISIC 2019 datasets, incorporating class balancing, noise removal, and data augmentation techniques. Learn more
Machine Learning Algorithms Implementation from Scratch: Machine learning algorithms are crucial in the data-driven world, enabling accurate predictions and automation. Implementing them from scratch fosters a comprehensive understanding, promoting creativity and customization. It enhances problem-solving skills, builds resilience, and empowers practitioners to interpret results effectively, making informed decisions based on the model's predictions. Overall, implementing algorithms from scratch cultivates technical expertise and a deep appreciation for their intricacies, driving innovation in the field of machine learning. GitHub Link