Last week, LeddarTech announced with partner Osram a long-term agreement for LeddarTech to provide LiDAR hardware and software components for Osram’s Percept LiDAR platform. Osram pitches Percept as the first flexible, solid-state LiDAR platform engineered with a strict focus on industrialization and automotive qualification.
In a Tier 2 role, Osram wants to offer a mid-range to long-range platform to OEMs, Tier 1s, and system integrators that can be adapted for specific applications. The partnership targets delivery of the industry’s first automotive-grade advanced driver-assist system (ADAS) and eventually fully autonomous driving (AD) systems at mass-market pricing.
“Early in 2017, Osram recognized LeddarTech’s promising technological LiDAR approach based on unique software and hardware components and therefore invested $70 million to become a major shareholder,” said Olaf Berlien, CEO of Osram Licht AG. “Since then, we have committed over $70 million more on developing LiDAR products, which has allowed Osram to offer LiDAR solutions to Tier 1s and OEMs that meet their cost and performance requirements for mass deployment.”
The platform integrates LeddarTech’s LeddarEngine, which comprises a family of system-on-chips and related LiDAR measurement software that are said to significantly reduce system cost and development time. Combined with Osram’s laser products, optical module design, and industrialization expertise, the Percept platform is claimed to be the most versatile and highest performance available at a cost that can enable volume deployment of ADAS systems integrating LiDAR.
The partners will also collaborate on perception software solutions that deliver an enhanced 3D environmental model based on LeddarTech’s perception technology including raw data sensor fusion. This technology further contributes to deliver enhanced and cost-efficient ADAS systems through the fusion of cost-efficient LiDAR, cameras, and radar and with lower overall system computing power.
Charles Boulanger, CEO of LeddarTech, believes that Osram’s presence selling into every major OEM with its LiDAR platform will transform the automotive industry and enable mass deployment of ADAS systems, leading to a significantly safer and vastly enhanced user experience.
“LeddarTech’s expertise in LiDAR, sensor fusion, and perception technology, combined with Osram’s industrialization competencies and their track record in delivering highly reliable and quality automotive modules, is destined to be the model of a winning partnership for the industry,” he said.
Boulanger also believes that Tier 1 automotive suppliers, supported by technology providers such as LeddarTech, will dominate the market for LiDAR sensing systems just as they have with radar and other sensing platforms.
Moving the bar in 2020
The Osram announcement follows significant growth in customer engagements, partnerships, and investment in 2020 for LeddarTech, which has been expanding beyond LiDAR to become a global leader in Level 1 through 5 ADAS and AD sensing technology. In November, the company was recognized by Tracxn in a category of only six Canadian company Unicorns, defined as one with a valuation exceeding the billion-dollar mark, representing the elite of the Canadian tech startup sector.
In 2020, LeddarTech reached over $350 million in investments from industry leaders and delivered over 9000 low-cost solid-state LiDAR sensors to customers, a double-digit increase over 2019. The company also announced volume production of the Leddar Pixell with manufacturing partner Faurecia-Clarion Malaysia.
Last year, Leddartech contracted with six Tier 1 and OEM customers to develop LiDAR measurement software, sensor fusion, and perception technology for ADAS and autonomous driving applications with a lifetime value of over $1.5 billion, supporting a growing opportunity funnel well over $4.0 billion. One of those OEMs is MILLA Group, which selected the Pixell as the front-end LiDAR for its POD autonomous shuttle. Another, Coast Autonomous, is the first deployment of the Pixell on an autonomous delivery vehicle at the Kinney County Railport near Brackettville, TX.
The company added four major global technology companies as collaborative partners for joint delivery of LiDAR solutions to the market within the Leddar Ecosystem including STMicroelectronics, Flex, dSPACE, and Ningbo Sunny Optical. It expanded collaboration with Renesas, combining its industry-leading raw data sensor fusion stack and LiDAR technology with Renesas’ newly launched R-Car V3U, a best-in-class ASIL D SoC (system-on-chip) for ADAS/AD systems.
The company also upgraded its automotive sensing solutions through two acquisitions.
VayaVision added a vital building block by combining sensor fusion and perception technology with the LeddarEngine platform. Built on open software architecture and combined with LeddarVision, the platform enables the company to address customer needs for sensing solutions that are hardware agnostic, scalable, and adaptable to any vehicle and sensor configuration.
The addition of Phantom Intelligence advanced LeddarTech’s strategy to aggregate and consolidate automotive sensing technologies, enabling the company to offer more comprehensive solutions to our customers at a lower cost.
Beyond LiDAR, all sensing
The acquisitions of VayaVision and Phantom Intelligence, combined with over a decade of expertise in Level 1 through 5 ADAS and AD sensing technologies, are meant to demonstrate the company’s commitment to continuous innovation and service to its customers.
Originally known for its LiDAR solutions, Frantz Saintellemy, President and COO of LeddarTech, said his company made the strategic decision to shift to “all sensing” in 2015.
“We come from the heritage of LIDAR because we’ve done that for well over 10 years now, but what we do now is a lot more,” he said. “We enable customers to build advanced driver assistance systems using camera/radar and camera/radar/LIDAR sensing and have a sensor agnostic [solution with] complete portability. So you can mix and match different types of sensors, and you don’t have to change your software every single time.”
From the outset, LeddarTech always wanted to be a sensing-platform company, said Saintellemy.
“We started with LIDAR, because that was what we researched, and we built modules based on that research.” However, the company’s leaders quickly realized that we had some key learnings that “helped shape our strategy in many ways.”
The end goal was had to be improved ADAS safety, so the company looked to improve existing camera- and radar-based solutions, while providing a seamless and extensible path to add LIDAR for more advanced autonomy.
“Today, [companies] are pushing LIDAR centric solutions not compatible with existing camera- [and] radar-based solutions, so they have to reinvent the wheel,” said Saintellemy. “We enable Tier 1s and OEMs to start with their camera/radar solutions, augment them, but also give them a path to upgrade adding LIDAR without having to redevelop completely.”
Beyond sensing, they realized the importance of open collaboration.
“You can’t have a closed approach,” he said. “You need to have a collaborative approach where everyone’s working towards the same goals together. You cannot look at ADAS or AD as a box; you have to look at it at a system level.”
Pushing raw data fusion
The industry has been pursuing object fusion, basically every sensor doing its own perception and all the perception data being fused in a central processor. That worked when there were fewer sensors.
“Every camera, every radar, they all have to have their own processor, plus you have to have another central processor,” said Saintellemy. “Multiply this by five or six cameras, by three or four radars, and then maybe you add on top of that LIDAR, you can see that it becomes a complex. Having each sensor node doing their own perception doesn’t scale well; it’s not a viable option.”
The more viable and safer solution is to employ raw data fusion, he says. Each sensor sends its raw data directly to a central processor, where it is combined or fused.
“Now, you have not only more efficiency, but you also speed up the processing, you reduce the error rate, you are taking the best out of each sensor,” he said.
The total system validates the scene and makes the best use of sensor redundancy, creating a much safer system.
“That’s what we’re seeing in the market,” said Saintellemy. “Over the next decade, the whole industry is about the shift from object fusion to raw data fusion.”
Providing actionable data
Modern cars generate lots of data, and that data can feed back into the development loop and improve the performance of the car at regional, city, and global levels. OEMs want to differentiate themselves, and access to that data will enable them to have better services to the end-user and can facilitate changes to improve their business models.
“Today, [OEMs] don’t have access to the data,” said Saintellemy. “Typical sensing companies supply a black box. That does not scale.”
LeddarTech is offering OEMs a solution that is open by nature so they can continue to evolve the services and the features that they can use to differentiate themselves to the end-user.
“In a nutshell, we’re changing the way people see sensors in the automotive space,” he said. “As opposed to being a hardware box or a sensor discussion, it’s about actionable data. It’s about having open, flexible, portable architectures that you can move from high-, to mid-, to low-end without always changing the architecture.”
LeddarTech is helping to coordinate this effort to provide actionable data with its Tier 1 and 2 partners.
“We’re all working to the same objective, which is essentially to enable actionable data at the lowest cost possible in as seamless a way as possible,” he said. “The Tier 1 does the assembly, while the OEM has the flexibility to use our platform to really differentiate their products across the geographies and across price points.”
