"Imagine that every day, once an hour, 24 times a day, a jumbo jet falls from the sky killing 350 people", says Yaron Toren, VP of Marketing at Oryx Vision. "This is the analogy for the number of people killed every day in traffic accidents worldwide. The cost is twofold: the direct cost of 1.3 million human lives lost and its' associated financial costs of billions of dollars. Obviously, the cause for these accidents varies from case to case, but 92 percent of all accidents occur due to human factors - a tired, drunk or simply careless driver."
According to Toren, this is the first and main reason that autonomous vehicles are the future, even if it takes time. Naturally, there are also other reasons for this development: the roads are becoming more crowded and people want to save the time they waste while driving. These three categories – saving human life and the utilization of space and of time – render the autonomous vehicle a major economic issue."
"The pace at which the world will make the transition to using autonomous vehicles is a matter of dispute. The bulls of Silicon Valley would say that it will happen within two months, however, in contrast to other industries, the automotive industry advances slowly and is not designed for frenetic progress. After all, millions of lives are at stake and any disturbance at the commercial or technological level causes inconvenience. In this industry, it takes years of testing for every small detail. It's no wonder then that when we look at what has happened to the motor vehicle since the 19th century, we see a very slow process."
"The transition to autonomous vehicles will be gradual. They will be initially limited to fixed routes in certain lanes on the road. However, a world in which at least most of the cars are driven by robots will nevertheless materialize. In any case, even when only a single autonomous vehicle takes to the road, it needs to perform properly and therefore, from a commercial perspective, it is already of interest today. More than 1.5 billion dollars has already been invested in various autonomous technologies."
"In order for a completely autonomous vehicle to be able to drive, it needs three capabilities. The first capability is to physically move the vehicle. It has already been proven that robots perform this function better than humans, however the vehicle also requires a second capability: decision making. This task, performed easily by the human driver, is difficult for a robot. A reasonable team of software engineers can overcome this challenge relatively easily 99 percent of the time, but an autonomous vehicle cannot take to the road with solutions for only 99 percent of the possible scenarios. The result would be huge numbers of people killed every year. Therefore, when constructing an autonomous vehicle, decisions must be correct with almost absolute probability – and that requires very extensive artificial intelligence capabilities."
"One of the key prerequisites, apart from the algorithms, is that the vehicle must have a precise perception of what is happening around it. The third required capability of an autonomous vehicle is therefore vision – it must be aware of what is happening in its immediate proximity. The major problem is in identifying each of the objects around the vehicle when some are moving and others are human, and there is need to predict their intended actions. This is one of the biggest difficulties in the development of an autonomous vehicle. Currently, there is no system that enables full vision at every speed, and in every kind of weather and visibility."
"This is one of the most bustling fields today and it's attracting many competitors worldwide", says Toren. "They are all proposing a range of methods that attempt to meet performance requirements." Two of these players are the Israeli companies Oryx Vision and AdaSky.
Like A Radio Antenna – Just with Light
The founders of Oryx Vision, Dudi Ben-Bassat and Rani Wellingstein, established the company with the objective of developing a unique machine vision technology for autonomous cars.
"There are different ways to give a vehicle eyes", Toren explains. "There are companies that maintain that a camera is the solution. But their picture is two-dimensional and immense calculating power is needed in order to create a three-dimensional image. Furthermore, the camera can't see at night and is confused by shadows. When radar is added to the camera, the precision of identification improves but still doesn't attain the required level."
"Most of the industry players assume that the vehicle needs an active vision system that sends a light beam, using the returning light to identify the objects in the surrounding area and their distance from the vehicle. These systems are called LiDAR (Light Detection and Ranging) and they do with light what radar does with radio waves. However, while this technology is used successfully in other fields, difficult problems arise when it is implemented on an autonomous vehicle and these are the problems that Oryx is attempting to solve."
"An autonomous vehicle needs to see for a distance of more than 200 meters in order to enable a safe stop on a highway. This sight must also be of sufficient resolution to identify the different objects. Furthermore, because we are referring to a vehicle that is moving, shaking and acquiring dirt, the system itself must be sturdy and stable. And if all that is not enough, complex systems come with a hefty price tag. In order to be available for public transport, the system's price needs to be reduced to mere hundreds of dollars."
"The existing systems we are familiar with cost tens of thousands, involve complex devices of revolving mirrors, and possess a range limited to only dozens of meters."
"There are two central elements in the LiDAR system – a laser that emits the light and a sensor that identifies the returning light. The laser beam hits the object and the pulses received chart a complete picture of objects and their distance. There are different methods for transmitting the laser beams, but all LiDAR systems are based on scanning and this is one of the system's major limitations."
"Oryx's major invention is that, using an industrial process, we have succeeded in creating a completely new component for light absorption. The light wave is part of the electromagnetic spectrum on which radio waves are also located. In principle, I can absorb and transmit light like radio – with antennas. The problem is that unlike radio waves, light waves are tiny and move at enormous speed. There are no electronics today that can contend with these speeds, but we have succeeded in creating a microscopic antenna that is capable of absorbing light and transforming it into an available electric signal."
"The system that we have developed enables detection at a signal-to-noise ratio a million times better than the other systems. We are translating this advantage into a flash system. Instead of scanning the scene, we distribute every pulse over the entire field of vision and receive a complete picture like a digital camera. This makes the system very simple, cheap and resilient. We are currently in the developmental stages and expect to be able to present a system that can be installed on a vehicle and provide high-quality output by next year."
The technology is relevant not only for an autonomous vehicle and for depth perception, but also for night vision, optic communications and, on a deeper level, computerization. We are not dealing with that presently but the technology is applicable to other industries."
Toren firmly believes in Israeli industry: "We saw with Oryx and with other companies that the Innovation Authority and other Israeli government and military entities are big supporters of innovation – beyond the grants and financial participation – and this is important in order to maintain the culture of Israeli entrepreneurship. If we approached a company in Denmark and asked who was interested in setting up a company, no-one would speak up. It’s just not part of the culture. In an Israeli high-tech company, everyone wants to participate."
The Sensor that Sees in the Dark
Another company benefitting from the financial grants and the support of the Innovation Authority is the startup AdaSky. Dror Meiri, the company's Vice President of Business Development says: "AdaSky's association with the Innovation Authority is less than two years old but during this period we have already received grants and other forms of assistance. In addition, we also work with the Fuel Choices and Smart Mobility Initiative and the Israel Export Institute."
"AdaSky was founded in January 2016 as a spin-off of a company that developed thermal vision solutions for the defense market. AdaSky CEO, Avi Katz, one of the owners of the parent company, had a vision that the technology they had developed would also be relevant for safety systems in autonomous vehicles."
"Autonomous vehicles must be capable of seeing and understanding the roads with complete precision in all weather and lighting conditions", Meiri explains. "Our technology aims to provide a solution to this."
"We offer broad solutions of thermal vision and vehicle safety that enables us to understand quite accurately what is happening around the vehicle, regardless of lighting conditions – including 'seeing' in the dark – even at large distances and high speeds and even when there are obstacles in the way such as boulders and blown-out tires on the road. The system we have developed is comprised of an infrared camera with FIR (Far Infrared) technology and complex algorithms of machine vision and artificial intelligence that process the necessary data and provide a complete picture in real time."
"The field of sensors is extremely bustling, both in Israel and abroad, and we are at the forefront of innovation. Our development team is probably the only one in the world today that processes thermal pictures for this market. The other existing solutions are excellent but do not enable an autonomous vehicle to operate in all conditions regardless of weather and lighting."
The company recently revealed its first product – Viper - defined as a groundbreaking solution that allows autonomous vehicles to travel at any hour and in any conditions. "Our solution is intended not only for autonomous driving but also for today's existing market of vehicle safety systems and smart vehicles", says Meiri. "It complements other sensing systems in scenarios where they don't function well – for example, in fog, haze, air pollution, rain and snow, and also in situations in which the conditions change such as tunnels, temporary blindness, dazzling, direct sunlight, sunrise or sunset and of course, in the dark."
"What happens on the physical level is that the camera gathers a thermal signal emitted from objects in the atmosphere and translates it into a high-resolution video signal. This signal is then analyzed and processed in order to make a decision about the recommended route and manner in which to drive. This process uses many complex algorithms and the signal is not a regular signal but rather, one that gives expression to the dissemination of heat from the object, therefore enabling us to observe both form and temperature. In this way, it is possible, for example, to distinguish between a live person and a statue or picture of a person etc. The technology can even enable us to determine whether or not the motor of another car we are looking at is in operation."
"Our product meets the highest safety standards. We are talking with customers about commencing car production in 2020, but we will actually be ready well before then", Meiri promises. "We don't have many competitors because our technology is extremely complex and requires especially high expertise. The people from our parent company have been working on it since before the turn of the century."
"Our vision is comprised of two objectives: becoming a dominant and influential player in the motor vehicle market and saving lives. It sounds pretentious but the motto guiding us is 'Driven to save lives' and we really mean it. Our system can significantly aid the attempts to lower the number of casualties from road accidents. Many victims are injured while driving at night, when visibility is poor. With the help of our system, it's possible to see, even hundreds of meters, in complete darkness. This represents a significant improvement that will have real influence."
Gil Shaki, Senior Director of Energy, Cleantech, Infrastructure & Transportation at the Innovation Authority: "The Innovation Authority is acting together with important government partners such as the Ministry of Transport and the Fuel Choices and Smart Mobility Initiative in order to fully realize the potential of the automotive and transportation revolution. The operational philosophy is based on three central layers: first, the establishment of a trials center that will mainly serve startup companies in need of trials infrastructures; second, a support program for pilot trial stages and for real-time demonstration of innovative technologies; and third, promotion of an innovation-supportive regulation that formalizes the conditions for performing an examination of innovative technology in an uncontrolled environment. This is a necessary stage in the fruition process of technology prior to its progress to commercialization stage.