Innovation Forums: Towards Autonomous Driving

  • The forum will be held at 16:00 - 18:00 on 03/28 (Thursday)
    • Moderator: Yu Huang
    Chief Scientist of autonomous driving and President of Silicon Valley Research & Innovation Center of Singulato Inc.
    Biography:

    Yu Huang is Chief Scientist of autonomous driving and President of Silicon Valley Research & Innovation Center of Singulato Inc. Before that, he had worked for Baidu R&D USA, Intel corporation headquarter at San Jose, Samsung R&D USA, Huawei R&D USA and Thomson Corporate Research USA. His main work fields are autonomous driving, computer vision, machine (deep) learning, VR & AR, image & video processing and cloud computing etc. Yu received his B Eng. from Xi’an Jiaotong University, M. Eng. from Xidian University and Ph D from Beijing Jiaotong University. He also was Postdoctoral Fellow at Tsinghua University, Alexander von Humboldt Fellow at U. of Erlangen-Nuremberg (Germany) and Postdoctoral Associate at University of Illinois at Urbana-Champaign (UIUC). Yu has published more than 30 academic papers in int. conferences and journals, filed more than 30 US/Europe patents and 13 of them issued.


    Hao Jiang
    Principal Researcher and Research Manager at Microsoft Cloud & AI
    Talk Title: Self-driving Car and Ambient Intelligence

    Abstract:
    Self-driving is becoming reality. With the help of Lidar, SLAM, computer vision and deep learning, a self-driving car can successfully maneuver on busy streets. However, there is very little information sharing among cars and what a car can see is still limited by the sensors' range. In this sense, self-driving is still quite dangerous. Ambient intelligence enables a smart environment, which can potentially solve the problem. By sensing the dynamic world using a large set of cameras and digitizing everything in the scene, we are able to provide self-driving cars the complete information to negotiate with different objects in the space. In this talk, I would like to discuss different computer vision methods I developed for self-driving, share what we are working on at Microsoft about ambient intelligence and propose how ambient schemes can improve the self-driving technology.
    Biography:

    Hao Jiang received PhD in computer science from Simon Fraser University in 2006. He was a Postdoctoral Research Fellow at the University of British Columbia in 2006-2007. From 2007 to 2017, he was an Assistant Professor and then a tenured Associate Professor in the Computer Science Department at Boston College. His research includes object tracking, matching, segmentation, human pose and action recognition, 3D computer vision, egocentric vision and deep learning. He developed the tracking algorithm for the Amazon Go, when he was a Senior Researcher at Amazon in 2013-2014. In 2016-2017, he led the Computer Vision Team at AutoX at Silicon Valley working towards pure vision approach for self-driving cars. Currently, he is with Microsoft working on robust computer vision to automate the everyday tasks.


    Junwei Bao
    Co-founder and CEO of Innovusion
    Talk Title: Image Grade LiDAR: Providing Sharp Eyes to Autonomous Vehicles

    Abstract:
    LiDAR has been generally viewed as an essential sensor for autonomous vehicles. During the past few years, a large number of LiDAR technologies have been proposed with wide variation in specifications. In this talk, we analyze the basic requirement in detection distance, angular resolution, and other aspects for LiDARs for Level 4+ autonomous driving, and demonstrate the quality and effect of high resolution LiDAR to perception and localization for autonomous driving. Different technology architectures are also compared for LiDAR design.
    Biography:

    Junwei Bao is currently the co-founder and CEO of Innovusion, an industry leading High Definition LiDAR developer for the autonomous vehicle and ADAS markets located in Silicon Valley. During his career, he has spent nearly 20 years developing innovative technologies and products in the precision instrument and optical sensor space. Prior to founding Innovusion, Junwei led the Sensor and Hardware Team for the Autonomous Driving Business Unit at Baidu. In other previous roles, he co-founded Timbre Technologies, Inc., the startup that invented and productized OCD, the “microscale LiDAR” for semiconductor manufacturing, and led Tokyo Electron’s metrology division in Silicon Valley. Junwei holds a B.S. in Physics from Peking University. He received his M.S. and PhD in Electrical Engineering at UC Berkeley. He has more than 40 patents in the area of optical metrology and semiconductor process control.


    Shiyu Song
    Principal architect and technical lead of Baidu Autonomous Driving Business Unit
    Talk Title: Towards Learning-based Localization for Autonomous Driving

    Abstract:
    We present a robust and precise localization system that achieves centimeter-level localization accuracy in varied city scenes. Our system adaptively uses information from complementary sensors such as GNSS, LiDAR, and IMU to achieve high localization accuracy and robustness in various challenging scenes, including urban downtown, highway, tunnel and so on. Moving forward, we introduce our latest work in exploring learning-based localization system. It leverages the help of the deep neural network structures to establish a learning-based approach that achieves centimeter-level localization accuracy, comparable to prior state-of-the-art systems with hand-crafted pipelines.
    Biography:

    Shiyu Song, PhD is the principal architect and the technical lead of the mapping and localization team at Baidu Autonomous Driving Technology Department (Baidu ADT). Before joined Baidu, he was a research scientist at NEC Labs America working on visual SLAM, vision based vehicle localization system for autonomous driving. He joined Baidu in 2014. He is one of the founding team members of the Baidu autonomous driving car project. His research interests include map reconstruction, structure from motion (SFM), SLAM and localization, etc.


    Anurag Ganguli
    Technical lead of perception and senior principal scientist of Plus.AI
    Talk Title: To Develop Autonomous Driving Technology of Commercial Trucks at PlusAI

    Abstract:
    PlusAI is developing a full-stack self-driving technology to enable large scale autonomous commercial fleets. Our truck fleets are currently being tested in multiple locations in the US and China. Building an autonomous truck, that is both safe and efficient, presents unique challenges across the different layers of the technology stack. In this talk, we will discuss some of these challenges along with how PlusAI is addressing them.
    Biography:

    Anurag Ganguli received his Ph. D. in Electrical and Computer Engineering from the University of Illinois at Urbana Champaign. He currently leads the perception team at PlusAI. Prior to joining PlusAI, Anurag was a research manager at the Palo Alto Research Center (PARC) working on a number of technologies for Intelligent Transportation Systems. His previous roles have included developing collision warning systems for cars and autonomous drones. His research interests include all things perception, optimization, machine learning, planning and control systems.


    Chang Yuan
    CEO & Co-founder of Foresight AI
    Talk Title: A Scalable Data Platform for Autonomous Driving at Foresight AI

    Abstract:
    In this presentation, I will introduce the technologies and product being developed at Foresight AI Inc. We are building a scalable data platform to empower the emerging mobile robots, such as robo-taxis, delivery trucks/drones, and flying cars. We have developed novel computer vision and machine learning technologies to generate accurate (~10cm), geo-referenced, three-dimensional, semantic data, including environment map and real-world dynamic scenarios. These data will guide and train the robots to navigate in the complex world and interact naturally with humans and other robots. We have also developed a vertically integrated solution with compact and low-cost 3D sensor kit and a scalable data processing pipeline. Both qualitative and quantitative results will be presented.
    Biography:

    Chang Yuan is the CEO & Co-founder of Foresight AI, an AI and robotics technology company. Chang received his Ph.D. from the University of Southern California, and B.Eng. from Tsinghua University, China, all in Computer Science. His expertise areas include computer vision, machine learning, robotics, heterogeneous computing, etc. His PhD thesis topic was airborne video based 3D world modeling and motion tracking. After graduation, he has been building cutting-edge AI and robotics technologies and products, including Apple Special Project, FaceID, Amazon Go, and Sharp 4K TVs. He has authored more than 15 papers and holds more than 10 patents.