Tech Convergence Will Spur Demand for New ADAS Technology

MVIS Software Push Leverages LIDAR Sensor Capabilities


With most of the focus and attention on the hardware capabilities of MVIS Dynamic View Lidar sensor, I thought it may be worthwhile to look at how MVIS looks to be seeking to leverage this hardware platform with proprietary software to create an integrated system for automotive safety and autonomy. 

From the Q1 2021 prepared remarks (April 29, 2021):

Sumit Sharma: "Having what I believe to be the best-in-class first generation sensor gives us a huge step up against competition. It also provides our very capable team with a hardware platform to further increase value for potential partners and our shareholders. In the short term, I expect our team to continue focusing on internal and external validation of our first-generation lidar sensor and any potential confidential evaluation from customers or partners. In the long-term, I believe a future sensor could provide features like Active Emergency Braking, Active Emergency Steering, Pedestrian Active Emergency Braking, and Active Lane Keep, among a longer list of higher level ADAS features with MicroVision software running on our edge computing. I believe a lidar sensor with embedded software that does not require massive amount of external computing will ultimately reduce cost of systems for OEMs, thus potentially accelerating adoption of vehicles with autonomous driving and active safety systems. I expect that key features in our first generation sensor like highest resolution, full velocity [output], immunity to sunlight and other lidar could allow an incredible opportunity for us to add significant value with our software for a greater sustainable strategic advantage. I believe future products built with our software, sensor performance, edge computing and scalability, would be valuable to OEMs, Tier 1 automotive suppliers, companies that are focused on mobility as a service and, therefore, of value to our shareholders."

Back in April 2020, I wrote a blog post on what I perceived to be Facebook's key weakness -- that it has no physical platform or OS and is dependent on Google Android and Apple iOS to bring it's "family of apps" to market. Recently we have seen how this risk manifests -- that when Apple makes a change to its terms of service, it can have serious negative impacts to Facebook's business. If Apple decided to kick Facebook out of the App Store, we would expect the impact to be acute. To address this vulnerability, Facebook is investing heavily to establish its own hardware platforms so that it has control of the experience end-to-end. So, just having software is not enough to establish real dominance. 

Likewise, providing hardware components without value-added software is limiting. To illustrate this point, the top 64 semiconductor companies are cumulatively worth $3.643T as of this writing. Whereas, the top 64 software companies (excluding Apple as it is vertically integrated with hardware and software together) are worth a combined $7.828T. So, software as an industry is worth more than double that of silicon hardware, probably because it has lower costs (no manufacturing) and it is what provides the direct value to the end user.

So we can see that the path to maximizing value creation (as Apple demonstrates and as Facebook aspires to) is to have your own software providing value and user experience on top of your own hardware. MVIS understands this and is playing offense by building a software capability in-house through aggressive hiring: 

A peek inside some of these job listings shows the capabilities that they are driving towards:

Sensor Fusion Algorithm Engineer:

Essential Responsibilities:

Develop algorithms for sensor fusion and object association between LiDAR and other sensor modalities.

Utilize fused sensor data to develop algorithms for Level 2+ and Level 3 Autonomous Driving functions.

Optimize algorithms for performance while minimizing computational complexity.

==

This role demonstrates MVIS' intention to provide an integrated sensor capability fusing inputs from LIDAR and other sensors such as 2D cameras or radar, to provide the most comprehensive and robust  understanding of conditions around the vehicle. This extends MVIS value proposition to be not just providing a LIDAR sensor, but the vehicle's real-time operating system. 


Design and development of algorithms utilizing MicroVision lidar's 3D point clouds as applied to Advanced Driver Assistant Systems (ADAS) and Autonomous Vehicles (AV).

Optimize algorithms for performance while minimizing computational complexity.

Analyze and process MicroVision 3D Point Cloud data sets for variety of operations, such as 
segmentation, clustering, collision avoidance, path prediction, etc.

Collaborate closely with software and RTL developers to enable implementation of algorithms in embedded software or logic.

==

These algorithms that analyze the 3D point cloud input can be embedded directly in MVIS silicon as edge computing that instruct the vehicle how to navigate to avoid obstacles or collisions based on the LIDAR sensor input. 

So, we can see that MVIS is gearing up to own not just the hardware side of the ADAS/Autonomous vehicle market, but to leverage that platform into high value embedded software. 

From the Q2 Conference Call transcript (August 4, 2021): Sumit Sharma: Today, our focus remains on establishing ourselves as a trusted partner for automotive OEMs, tier ones, and mobility-as-a-service companies. We expect our closely integrated hardware and software to enable the ADAS safety market while also supporting potential partners in the autonomous driving market. We expect to start validating our ADAS safety capabilities to our track testing program by second-quarter 2022.

Nothing is more important to us than getting this right with our products and partnerships. Our technology and company have already demonstrated our pedigree and enabled large global brands. I expect us to establish ourselves and lead the ADAS and autonomous driving space with hardware and software into the future. We remain confident in our ability to successfully execute on our strategy.



Comments

  1. Great analysis Ben. I do question whether automotive OEMs are focused right now on improving their ADAS systems or really targeting investments in LiDAR for their L3-L5 solutions. It's gotten to a point where I feel most consumers appreciate the current capability set of their ADAS-equipped vehicles and aren't really expecting any incremental advances.

    For example in my 2018 Volvo XC-90, I have lane-keep, emergency braking, and adaptive cruise control. These features could be better but they're "good enough" that they're quite usable already, especially when conditions are optimal (e.g., Highway driving, long distances). This is in a now 3 year old car, and my point is part of the reason why I don't think OEMs have made major investments in LiDAR for their ADAS systems is in large part because consumers aren't demanding it for ADAS because the existing ADAS solutions are "good enough"

    Just a thought, and would love your perspective on this. Been buying and holding MVIS for over a year and no plans to sell before a major customer / partnership is announced.

    ReplyDelete
    Replies
    1. I believe there's an arms race for vehicle autonomous capabilities -- and many new entrants are coming into the market piggybacking on electric vehicles. Couple that with the fact that there has never been a LIDAR sensor with the resolution/frame rate, form factor, price scalability and capabilities as the MVIS sensor -- expected to be available for sale this quarter. So, we will see what happens with that. I expect that autonomy is the differentiator with these new OEMs (rather than, say, plusher leather).

      From the interview with Joanna Makris: After all, lidar appears to be the go-to technology for carmakers looking to enable more intelligent driving. As Sharma states, “Every OEM and every Tier One that was present [at the Munich Auto Show] had a lidar story as part of their product story.”

      Delete
  2. We’ve always had good hardware, i.e. pico projection and LiDAR, so are you saying that we’ve never had the software development to add value? I’m not sure – I’ve seen great demos for pico, interactive projection and LiDAR! but still no takers.
    Several years ago I talked with a Microsoft AR developer who had hired several Microvision engineers and he was not impressed. Currently, I worry that this hiring may be too little too late. Where are we going to find experienced engineers in this tight labor market?

    ReplyDelete
    Replies
    1. "hired several Microvision engineers and he was not impressed". So he hired them because they were unimpressive?

      Delete

Post a Comment