Sample Projects

PS: The links below are being provided as a convenience and for informational purposes only by me in my personal capacity. Only articles that are published in a scientific or professional journal, or editorial, or service for a scientific or professional journal are presented in this page.

Please read the Disclaimer for additional details.

NIO Autonomous Driving

Improved localization using visual features and maps for Autonomous Cars, 2018 IEEE Intelligent Vehicles Symposium (IV), [link]

Additional info: Vision-Guided Self-Localization for Autonomous Cars, MATLAB Expo 2017, Invited Talk

Robust method for removing dynamic objects from point clouds, 2020 IEEE International Conference on Robotics and Automation (ICRA) [paper]

Motion based calibration between multiple LiDARs and INS with rigid body constraint on vehicle platform, 2020 IEEE Intelligent Vehicles Symposium. Sample 3D point cloud data shown in the following image and also in the video above. [paper]

Cutting-edge fusion of 3D Point Cloud Data from 5 LiDARs, ensuring impeccable quality. Auto-calibration on-the-go while the vehicle cruises the road.

Toyota InfoTechnology Center, U.S.A., Inc, Intelligent Computing Division

(now Toyota InfoTech Labs)

CHASE Algorithm: "Ease of Driving" Classification

Abstract:

In this paper, we present the concept of "Ease of Driving" or EoD rating for road scene environment. We introduce the Classification by Holistic Analysis of Scene Environment (CHASE) algorithm to automatically classify the EoD rating using computer vision and machine learning techniques. With the CHASE algorithm we were able to achieve an accuracy of 86.84 % to classify the EoD. In the near future, EoD information will be used to create better navigation services to improve the driving comfort and mobility by leveraging the existing camera sensors in the car. [paper]