Sign up for our weekly emails to stay up to date on the latest news!
In an exclusive interview with Self Drive News, Eran Ofir, CEO of Imagry, shares his vision for the future of autonomous driving. With a rich background in leading global hi-tech companies across four continents, Eran has a proven track record of driving business growth and facilitating major acquisitions by industry giants like Huawei, NEC, and NCR. Now at the helm of Imagry, Eran is spearheading innovative solutions that address critical challenges in autonomous vehicle technology. From enhancing safety and convenience to overcoming the global bus driver shortage, Imagry’s HD-mapless and bio-inspired approach is poised to revolutionize the transportation sector.
What specific problems is Imagry addressing with its autonomous vehicle technology?
Safety and convenience are some of the biggest factors spurring the introduction of autonomous driving capabilities in passenger vehicles. AI-based solutions can reduce accidents by eliminating human error. Relieving the driver of tedious tasks with applications such as cruise control, traffic jam assist, automatic braking, etc. will make the trip more enjoyable.
For the public transportation segment, autonomous driving addresses the bus driver shortage (currently at about 15% worldwide, and growing) by enabling Public Transportation Operators to add to their fleets to increase route frequency, reduce operational costs, and extend hours of operation without hiring more drivers.
Why do you believe that HD-Mapless is the right approach going forward with autonomous driving?
The advancement of self-driving technology has often emphasized the importance of HD maps. However, Imagry posits that autonomous navigation doesn’t require these detailed maps, which are expensive to purchase and maintain. Instead, our technology allows vehicles to drive independently by perceiving and mapping the environment in real-time, similar to how a human driver behaves. This approach, shared by Tesla, promotes vehicle autonomy without the need for constant map updates, reducing communication costs and cybersecurity risks. Imagry’s onboard computing platform processes all necessary data for safe navigation and periodically receives updates “Over The Air” to enhance its AI-driven system. Imagry envisions a future where AI’s adaptability and dynamism, not HD maps, will guide autonomous driving.
How is Imagry’s autonomous driving solution different from ADAS solutions?
Both these autonomous driving approaches utilize sensors, hardware, processors, cameras, etc., but their integration and computing power requirements are different. ADAS (Advanced Driver-Assistance Systems) are a collection of applications (e.g., cruise control, lane keeping assist, automatic braking, etc.) that can help drivers operate their vehicle safely and confidently. While ADAS developers believed they could achieve SAE L3/L4 autonomy by patching all these applications together, we took the opposite approach, which was proven to be the one accepted by the market. We developed from the get-go software that would support SAE L4 autonomy, where our SAE L3/L4 products roadmap was based on two restrictive factors: computing power and regulation. Following Moore’s Law, computing power is rising exponentially. Regulation for autonomous driving is currently being formulated in many countries around the world.
What do you mean when you say that Imagry’s solution is “bio-inspired”?
We decided that the best way to create a safe and reliable autonomous driving solution is to use the latest AI technologies to mimic the behavior of a skilled human driver. Humans learn via a combination of observation, imitation, and practice. We have designed a system that works on the same principles. Our software solution is composed of two software modules: the perception layer (like our eyes) and the motion planning layer (like the cortex in our brain). The perception layer utilizes an array of cameras, positioned around the vehicle, to create a 360° representation of everything in the near vicinity, up to 300 meters. AI technology, in the form of deep convolutional neural networks (DCNN), identifies each object and its trajectory, and then passes this map to the motion planning layer which determines the correct course of action: steering direction, braking, and acceleration.
Imagry’s software solution enables the autonomous vehicle to understand the road while driving and react to dynamic contexts and environments, just like an experienced human driver; this is the source of the term “bio-inspired”.
What is unique about Imagry’s solution?
There are four main factors that make the Imagry solution unique.
1) HD-Mapless: It bypasses the use of HD-maps, which are expensive to purchase and incur continuing communication costs through the cloud (which poses an increased cyber risk and assumes stability in network coverage). Instead, Imagry technology relies on an AI-based perception layer (like Tesla does) to navigate the current environment in real time.
2) Hardware Agnostic: It is hardware agnostic for both sensors and computing components. While other autonomous driving software solutions dictate the use of specific hardware, Imagry allows OEMs and Tier-1s to determine which components to use, commensurate with the vehicle and model under design. This feature reduces integration timelines, keeps costs down, and most importantly – provides the agility and flexibility that OEMs need in their Software Defined Vehicles.
3) Location Independent: Another benefit of the HD-mapless approach and the distributed neural networks architecture (as opposed to the one-big-neural-network approach) is that the solution supports a scalable rollout to locations globally. It also supports both left-and right-hand drive, which is a quality in the market that is unique to Imagry.
4) Supervised Learning: Unlike the Unsupervised Learning method used by Tesla, wherein all data is dumped into one large neural network which then dictates the behavior of the vehicle, the Imagry solution is trained using our proprietary annotation tool. Although this method demands more resources in the short term, the result is safer, law-abiding road behavior based on existing road infrastructure.
Why has Imagry chosen to focus on autonomous passenger cars and buses rather than other forms of autonomous transportation?
Imagry’s focus has always been creating a good product-market-fit for the sectors in which it works. Currently in the market, technological autonomous driving solutions that were built for robotaxis are not suitable for passenger vehicles as they are much too expensive. Imagry’s cost-effective solution was tailored to fit the required cost structure for inclusion in mass market passenger vehicle.
For autonomous buses Imagry uses the same software that is being used in the passenger vehicles, but with more powerful computing required to support L4 autonomy. The autonomous bus market is escalating now across the globe, given the mounting pressure on public transportation due to the aging of populations, and urbanization, coupled with the shortage of drivers required to meet those demands.
Imagry has acquired a unique position in the market, proving that its award-winning and approved technology fits a number of use-cases that are both long term (passenger vehicles) and shorter term (autonomous buses).
Is the Imagry autonomous bus solution field-proven?
In 2023 Imagry was awarded three autonomous bus projects in Israel. Two of them are already on the road, one in the largest medical center in the Middle East, and the other is a commercial service line on public roads in the city of Nahariya. In these two projects, the buses are operated by Nateev Express, the third-largest public transportation operator (PTO) in Israel.
Imagry was also selected by TransDev (the world’s largest PTO, with service in 21 countries) as an Autonomous Transport System (ATS) certified vendor. We will start working with them in Europe this summer. Later this year Imagry expects to win two additional autonomous bus projects with other PTOs, one in Europe and the other in APAC.
Why do you think that the time has come for autonomous driving, after so many promises we have heard over the last decade?
The two main dependencies for reliable and safe autonomous driving are computing power and regulation. We have seen growing computing power (as per Moore’s law that is being proven now also in the automotive domain) as well as progressive regulation towards L3/L4 autonomous driving in many countries around the world.
Also, we have seen consolidation in the market. Companies which didn’t take the AI route toward an autonomous driving solution have disappeared, and the same is true for companies that developed solutions that were too expensive for the target segments. Imagry has been driving autonomously for nearly five years with a growing fleet on public roads in the U.S., Germany, Japan, and Israel. This was needed in order to train the neural networks that are at the heart of Imagry’s AI-based solution.
We believe that urban mobility in 2030 will be a hybrid of autonomous and human-driven vehicles. We will see autonomous buses which will improve the service, safety, and frequency of buses for the public’s benefit, as well as quite a few autonomous passenger vehicles driving on public roads. Many of the new passenger vehicles sold by then will be ‘Software Defined Vehicles’ meaning they would be able to acquire new autonomous driving capabilities via over-the-air software updates. This will make the urban environment safer and more convenient.
How does Imagry align with the era of the Software Defined Vehicle?
Imagry’s L3/L4 autonomous driving solution is economical, easily deployable, and scalable. It can be deployed in its entirety, supporting a Software Defined Vehicle (SDV) architecture, or as specific software modules based on the use case.
Imagry’s alignment with SDV platforms was simplified because it is hardware-agnostic and can thus connect to a variety of computing platforms. When working with Tier-1s and OEMs, Imagry also provides autonomous driving ‘applications’ that run on top of the SDV platform. An example is the autonomous parking product that was developed with Continental. It allows the vehicle to explore a parking lot, detect a suitable parking space, and park the vehicle – all without the driver touching the steering wheel, or getting any help or data from resources external to the vehicle.
Other ‘applications’ that Imagry is offering are ‘Traffic Jam Assist’ for driving in a city, and ‘Safe Driver Overwatch™’ for young drivers and elderly ones.
Can you tell us a bit about the company?
Imagry was founded in 2015 for high-end computer vision applications and has specialized in the automotive sector since 2018. The company’s bio-inspired technology combines a real-time vision-based perception network and imitative artificial intelligence (AI) for a driving decision-making network that is economical, easy to deploy, and scalable. This eliminates the need for external digital mapping of roads and cities, and can be installed in passenger cars as well as buses.
Imagry has two main development offices – in San Jose, CA and Haifa, Israel. The company sells its products to Tier-1s, OEMs and Public Transportation Operators.
Sign up for our weekly emails to stay up to date on the latest news!
