LeddarTech Incoming CEO Saintellemy on Autonomous Driving Software

Exec-Edge

December 4, 2023 7 min read

By Exec Edge Editorial Staff

Exec Edge spoke with Frantz Saintellemy, Incoming CEO of LeddarTech, which is going public via a merger with Prospector Capital Corp. (Nasdaq: PRSR).

ExecEdge: Welcome, Frantz. Let’s start off by learning a little more about LeddarTech. Can you tell me more about the company and what it provides?

Frantz Saintellemy: Of course. Our mission is to improve the safety and quality of life for travelers, commuters, workers and mobility industry professionals by enabling applications that reduce traffic congestion, minimize the risk of road accidents and improve the overall efficiency of road transport. We offer a very crucial piece of the software stack for Advanced Driver Assistance Systems and Autonomous Driving Systems which provides vehicles with better environmental understanding models (that is, the 3D view of the vehicle’s surroundings), which is at the base of making a vehicle intelligent. Our software-only solution seeks to solve the limitation in currently used systems that are unable to scale up efficiently to the level of awareness required for new regulations and consumers’ expectations.

Our approach also helps reduce OEM’s development and system costs while reducing maintenance overhead withleaner architecture and savings on lower sensor and processor costs. These are critical, unsolved problems that OEMs face today. Our answer to these challenges is our innovative environmental sensing software solution which uses AI-based “low-level” sensor fusion and perception. The software platform is sensor and processor hardware agnostic, scalable and outperforms current object-fusion based architectures. This fundamentally different software architecture recreates the perceived environment around the vehicle, empowering OEMs to use fewer sensors and/or more cost-effective sensors. In basic entry-level trims, LeddarTech delivers nearly twice the performance at nearly 30-40% lower sensor cost of current “object fusion” alternatives adopted today, while delivering significantly more performance and cost savings in medium and premium trims.

ExecEdge: For those who aren’t as familiar with your technology or services, can you talk about your market and why investors should be paying attention now?

Frantz Saintellemy: The automotive industry is at a critical inflection point. OEM’s and Tier 1 /2 suppliers are equipping vehicles with more sensors than ever before – with the promise of increased safety through ADAS applications leading to fully autonomous driving in the future. The industry has embarked on a transition to “software defined vehicles” which has OEM’s thinking “software first” for better integrated, cost efficient, flexible vehicle architectures that are easier to maintain with continuous software updates. These OEMs are looking to deliver on that promise with an emphasis on safety, reliability and increasing consumer demand for convenient features that work reliably all the time –while keeping costs in check.

LeddarTech offers a scalable, cost-effective, sensor-agnostic software-only solution named LeddarVision that delivers highly accurate 3D environmental models and perception for vehicle manufacturers. Based on LeddarTech’s comprehensive and demonstrated expertise, LeddarVision processes sensor data centrally that comes from each sensor such as cameras and radars before filtering, to efficiently achieve a more reliable understanding of the vehicle’s environment required for navigation decision making and safer driving. LeddarVision resolves many limitations of ADAS architectures based on legacy object-level fusion by providing:

Scalability to AD/HAD (automated driving/highly automated driving)
Flexible modularity to effectively handle a growing variety of use cases, features and sensor sets
Centralized, multi-modality (meaning multiple different kinds of sensors) sensor-agnostic low-level fusion optimally fuses all types and combinations of sensors, such as radar, sonar, camera, GPS and Lidar as needed based on the features and level of autonomy required.
Low-level sensor fusion utilizes all the raw information output from each sensor without filtering at the sensor level for better and more reliable operation. As a result, this low-level sensor data fusion and perception solution provides superior performance, surpassing object-level fusion limitations in adverse scenarios like occluded objects, object separation, camera/radar false alarms, blinding light (e.g., sun, tunnel) or distance/heading estimation

The technology enables auto manufacturers and suppliers to solve critical sensing, fusion and perception challenges they are faced with to deploy higher performance systems. LeddarVision significantly lowers cost, increases performance and shortens time-to-market for customers through a unified software platform that leverages and improves previously trained algorithms. It also:

Allows the vehicle to better understand its environment, permitting better driving decisions.
Combines multiple sensor inputs into a single higher quality 3D model.
Delivers nearly 2x performance at nearly 30-40% lower sensor cost of current “object fusion” alternatives adopted today.
Enables broad sensor and processor sourcing opportunities for OEMs.
Allows OEMs to change their sensor suite without major software recoding or AI retraining.

ExecEdge: How do you differentiate from your competitors?

Frantz Saintellemy: The current ADAS solutions based on vision-centric object level fusion have reached their limits.They have:

Limited performance as separate integrated perception engines to each sensor detect and classify the scene; this filtered output is then fused with other filtered data from other sensors.
Filtered information that is lost and cannot be used in the fused environmental model to improve understanding of the vehicle surroundings. High cost and complexity for hardware, software and processing as well as costs associated with architecture (multiple software stacks to maintain, train, verify, validate and certify).
Hardware dependence of the software, which means any change in sensor position or characteristics requires significant perception software re-coding, training and testing to achieve required performance and safety.
A complete lack of scalability as architecture does not scale as features, sensors and sensor position vary across brands and models.

Our approach, AI-based low-level (also known as raw-data) sensor fusion, which represents the next generation ADAS software as OEMs seek increased performance and move away from object fusion is superior because it delivers:

High perception performance with nearly 2x detection range and greater reliability, as each sensor complements each other’s strength, also resulting in less degradation in difficult weather conditions and improved safety.
Lower computing & maintenance costs from leaner architecture and savings on sensors and processors, nearly 50% on current entry-level systems sensor costs.
Hardware agnostic solution provides go-to-market flexibility and shorter development cycles for OEM’s. Demonstrated 6-week integration of new camera and 4D RADAR.
Improved scalability from integrating with multiple sensors, resulting in a single software platform supporting multiple brands and models.

LeddarTech also reduces costs by breaking the requirement to take hardware in order to obtain access to the software. The tech:

Disaggregates hardware from software enabling OEM sourcing flexibility.
Allows OEMs to change or upgrade sensors without major recoding and AI retraining of their fusion and perception software.
Facilitates “over-the-air” and “backwards compatible” upgrades to meet NCAP / GSR2022 and other regulatory requirements.
Efficiently solves the fusion problem for OEMs as the number of sensors proliferates in the vehicle.
AI-based Low-level fusion technology enables customers to use fewer and lower priced sensors in their design while achieving higher performance

ExecEdge: Why should investors believe in LeddarTech?

Frantz Saintellemy: We believe LeddarTech represents a compelling investment opportunity for many reasons. Specifically, the company has:

A large, and fast-growing total ADAS and AD market that is expected to be $42 billion by 2030 with proprietary value-added solutions and a flexible business model leveraging the benefits of AI-based low-level fusion. We believe that the industry is increasingly moving towards low-level sensor fusion and perception as the solution of choice for next generation ADAS and AD systems. The market is driven by regulatory frameworks for collision avoidance and the resulting near-term need to unify data from the proliferation of sensors as ADAS systems evolve and increase performance requirements.
A disruptive sensor fusion and perception software solution that differentiates itself in the market through the use of AI-based low-level fusion instead of legacy object level fusion, which solves many fundamental industry challenges and has the support of various OEMs and Tier 1 & 2 go-to-market partners. The tech provides high-performance, cost competitive, scalable, flexible and sensor and processor hardware agnostic solutions that support ADAS and AD applications.
In our view, one of the few software-only investment opportunities to participate in the growing automotive software market driven by the proliferation of software-defined vehicles.
Early mover advantage, with a flexible business model that is easily scalable and offers opportunities for attractive software margins supported by a foundational and increasing IP portfolio (150+ patent applications).
An attractive valuation compared to market comparisons that are hardware-centric businesses compared to LeddarTech’s higher margin software business.