Get the latest on imaging and auto sensors, autonomous vehicles, and AI and more.
Focusing on the theme, The Road to Autonomy and Everything In Between, this year's program will feature 2 days of sessions and panel discussions addressing challenges from L2-L5 autonomy.
This session will provide an update on the current progress of autonomous driving and provide projections for the future of the industry identifying challenges and trends.
Junli Gu | VP of Autonomous Driving of XMotors.ai
Autonomous driving has gained enormous attention and momentum in recent years, due to its huge impact on transportation industries. This talk will share the perspective of XMotors, including strategies for massive deployment of self-driving solution, key insights on artificial intelligence evolution, personalized and geographically differentiated driving features.
Ronald Mueller | CEO of Vision Markets of Associate Consultant of Smithers Apex
Nazre Batool | Development Engineer Automotive Transport Solutions Pre-Development & Research of Scania
At Scania, our heavy-duty autonomous vehicles have endeavored to navigate in a variety of environments, from isolated mines to urban roads, autonomously. The adaptation of contemporary sensing technologies and sensor data algorithms to match our vehicles’ specific needs to reach these autonomous driving milestones has not been easy. The talk will include Scania’s experiences and specific challenges faced in this endeavor.
Panelists to be confirmed
Improving vision systems precision and performance, challenges from L1-L5
Haran Thanigasalam | Chair of Imaging and Camera Work Group of MIPI Senior Platform Architect and Independent Contributor
MIPI CSI-2 is rapidly advancing to meet the needs of emerging imaging and vision applications targeting multiple platforms including automotive. This presentation will focus on CSI-2's key features, capabilities, and select use cases mapped to product platforms. In addition, highlight new features presently in development for EOY 2018 and EOY 2019 releases.
Lauren Guajardo | Field Applications Engineering Leader of Teradyne
When the next sensor design is coming out, every advantage is needed to bring the device to market flawlessly. Are opportunities for savings left on the table because testing is overlooked? Device testing is necessary, but it can also be overly complex and resource draining if not met with a strategic plan. Understanding when to use automated test equipment (ATE) instead of bench testing equipment can reduce time to market and offer cost of test savings. This presentation will share case studies that demonstrate the advantage of ATE by providing faster and increased test coverage as well as early prototyping feedback to designers. Prototype testing with ATE When wafer probing is faster than the bench How to increase test coverage at probe Is your bench FPGA masking failures. What to consider when evaluating ATE and cost of test.
Abhay Rai | Head of Automotive Sensor Marketing of Sony Electronics
"Safety Cocoon" is a name expressing the creation of an area surrounding the car where a car can detect its surroundings in 360 degrees and prepare for danger avoidance from an early stage in various driving situations, enhancing safety in our society of constantly improving levels of autonomous driving. Adopting Sony's image sensor will contribute to achieving enhanced safety efficiently. The "Safety Cocoon" concept is explained by comparing the visibility of human, car and image sensor.
Presentation by Texas Instruments
Kenneth Funes Mora | CEO and Co-founder of Eyeware
Virtual co-pilots will be required to sense the driver's attention, for monitoring purposes and controlled handovers, to minimize human error. However, they will also use the eye gaze to allow a more natural interaction with in-cabin infotainment solutions.
Eyeware is proposing a new approach to solving the eye tracking problem using autograde depth cameras.
Sravan Puttagunta | CEO and Co-founder of Civil Maps
The next evolution in AV mapping and localization will occur at the chip-level as developers move from expensive, power-hungry systems to alternative architectures that are widely available and deployable with minimal cost and power consumption. This gives rise to the techniques involved in Edge Mapping, the real-time ability to create, update, and share HD maps in-vehicle.Edge Mapping pioneer Civil Maps will share how it has integrated with Arm’s architecture, licensor to the largest ecosystem of auto electronic systems, to achieve these goals. The company will also present road test scenarios such as urban, rural, tunnel, off ramp, and highway, where their edge-based approach was applied.
Sumanth Chennupati | Computer Vision Systems Engineer R&D of Valeo North America
Deep Learning offers solutions to highly complex visual challenges like classification, localization, and segmentation. Extraordinary performances on semantic segmentation, object detection etc., provides autonomous vehicles critical information about drivable areas, surrounding obstacles’ location and speed etc. The majority of the research has been done using single camera images. However, learning from multiple cameras to understand the surrounding will offer advantages in some critical cases. This presentation will review
Patrice Roulet Fontani | Vice President,Technology and Co-Founder of ImmerVision
Assisted and self-driving systems have generally reached the point of being able to analyse their environments and process decisions but are held back by the lack of adequate perception. The move from driver convenience to ADAS and full autonomy requires a truly Intelligent Vision system that includes perception suitable to humans and machines, and processes data captured through sensor fusion. The system properties must include low-latency, reliability, low-power consumption, and cross-platform support. High-resolution, wide-angle lens technology and tunable image processing, combined with sensor fusion technology, addresses capturing, channeling, synchronizing, and contextualizing the data from the multiple sensors. In this talk, we will explore the key roles of reliable, AI-classifiable, and actionable pixels and the lens and sensor technology required to capture them.
What challenges must the industry overcome from each end of the supply chain to reach level 5 autonomy.
Aditya Srinivasan | Innoviz of General Manager, North America
Fully autonomous vehicles are coming in the not-so-distant future, and it won’t be long before self-driving cars are available to everyone. But what will it take for autonomous driving to go mainstream? What technology is necessary to make it happen? How can the industry make prices reasonable for the masses? What obstacles stand in our way, and how can we overcome them? In this session, Innoviz CEO, Omer Keilaf, will answer these questions as others as he presents a roadmap for achieving mass commercialization of autonomous vehicles.
Rudy Burger | Managing Partner of Woodside Capital Partners
The impact of recent accidents and their effect on the market.
Speaker to be confirmed
This presentation will explore the impact of recent accidents on the autonomous driving market.
What challenges are there and how are they being met. Sensor performance and improvement, needs challenges
Dr. Daniel Winters | CEO of Trioptics
Today's automotive sensors like cameras or LiDAR systems have to work under stringent environmental conditions like e.g. high and low temperatures or strong vibrations. As the industry moves to more sophisticated sensors with higher resolution, error budgets during assembly become smaller, sensors become more sensitive to misalignment and need more accurate calibration.
State-of-the-art active alignment technology helps overcome these challenges by aligning optical elements and sensors individually in up to 6 degrees of freedom with micron-level precision. In combination with precise calibration this improves production yields and gives more consistent and better quality than conventional assembly techniques.
Anand Gopalan | CTO of Velodyne LiDAR, Inc.
Abstract to come
Dr. Allan Steinhardt | Chief Scientist of AEye, Inc.
The industry is shifting from raster scanning LiDAR technologies to intelligent vision systems that dynamically perceive a scene. First generation LiDAR systems, because of siloed sensors and rigid data collection methods, tend to oversample or undersample information. This requires significant processing power and time to extract critical objects, leading to latency. New systems are emerging that fuse intelligence with the data collection process - enabling the dynamic tracking of targets and objects of interest, with almost no computational penalty. AEye will delve into this new form of intelligent data collection, called LiDAR”, and its role in delivering human-like perception. Highlight the differences between legacy LiDAR systems and emerging intelligent vision systems - Discuss the differences in data collection methods - Explain how dynamic tracking works - Discuss LiDAR’s role in optimizing data collection, reducing bandwidth, improving vision perception and intelligence, and speeding up motion planning for autonomous vehicles.
Dr. Tomoko Ohtsuki | Product Line Manager of Lumentum
A variety of diode laser illuminators have been adopted in different LiDAR systems to cover demanding performance requirements. There are two types of diode lasers: vertical-cavity surface-emitting laser (VCSEL), and edge emitters. Each of these lasers has characteristics which are ideal for specific tasks. Diode laser illuminators of both types, developed by Lumentum over the last several years, have been adopted and advanced in the consumer electronics market. Reliability, and the ability to manufacture in volume with high quality are critical for the demanding consumer and automotive markets. We will present advancement of high power VCSEL array illuminators, manufactured in large volume production lines. Test results demonstrate these illuminators can reliably operate in the temperature range -40C to 115C, and under other environmental stresses required for automotive-qualified products. For an edge-emitting laser, we will introduce a new 1.5um tunable Distributed Bragg Reflector (DBR) laser prototype which will enable frequency modulated continuous wave (FMCW) Coherent LiDAR for >150m range.
Carl Jackson | CTO and Founder of SensL Division, OnSemi
The fusion of radar, LiDAR and camera data is a key trend for the future of autonomous vehicles. ON Semiconductor, through the acquisition of SensL Technologies, is now able to provide a LiDAR solution capable of long distance ranging with low reflectivity targets, through the use of their SPAD and SiPM technology. This combined with ON Semiconductors radar and image sensor expertise and experience in the automotive market, can be leveraged to co-develop the technologies from the outset to facilitate the fusion of sensor data. This presentation will highlight the ON Semiconductor vision for this fusion, leading to improved capabilities for the automotive market.
Full vehicle autonomy will be achieved through sensor fusion - ON Semiconductor's acquisition of SensL gives it the distinctive position of being able to provide solutions for radar, image sensors and now LiDAR for this market. - The co-development of the technology for these modalities will allow for a more robust fusion of the data.
Ram Machness, VP Product & Customer Succes, Arbe Robotics
The autonomous driving industry requires a reliable sensor that performs in all lighting and weather conditions. In addition, in the crowded autonomous environment where cars may drive one towards the other on highways at high speeds, next to bike riders or stationary objects, the sensor must be able to “see” them coming from 300 meters away, separate between them in real time, track velocity, and measure distance.
In this presentation, Ram Machness, will explain how a 4D imaging radar can overcome these challenges. He will focus on the technical aspects of achieving low false alarm rates, coping with mutual radar interference, providing high refresh rates, while keeping prices low and accuracy high.
Panelists include AEye, ImmerVision, and SensL