The Israeli start-up INTSITE seeks to continue the innovation revolution in advanced driver assistance systems, begun by companies such as Mobileye and Innoviz, and is bringing these advancements to heavy mechanical-engineering equipment, with an emphasis on construction sites and mines.
The company was founded in early 2018 by twin brothers Tzach and Mor Ram-On, 36, the former a civil engineer and the latter an aerospace engineer who previously worked at Rafael Advanced Defense Systems. The ADAS systems developed by the company (Advanced Driver-Assistance Systems) are based on image processing using artificial intelligence (AI) to analyze camera data in real time for heavy mechanical-engineering equipment, such as excavators, bulldozers and various trucks.
To date, INTSITE has raised $2 million in a seed round and recently closed a financing round for a similar amount.In contrast to companies that embark on large funding rounds, INTSITE has so far remained under the radar of capital-raising due to a creative business model that allows it to generate positive cash flow and not rely on large-scale external financing.
The company recently revealed its successful collaboration with Japan’s Komatsu, the world’s second-largest heavy-duty mechanical-engineering equipment manufacturer after Caterpillar.
The collaboration is reflected in five projects that have made Komatsu INTSITE’s largest customer. In addition, INTSITE is currently negotiating with Komatsu, which is traded on the Tokyo Stock Exchange at a valuation of $23 billion, to sign a commercial agreement whereby INTSITE’s systems will be embedded in the heavy mechanical-engineering equipment that Komatsu produces.
“We work with several types of customers,” explains Tzach. “There are end-users who suffer from low output of the equipment and safety problems at the construction site. This includes companies in construction and mining, and companies that use heavy equipment to create infrastructure. We contract with these companies through distributors, and we work with companies that rent the equipment to them. They offer their customers our company’s solution – the computer with all of its intelligence and the software in the cloud.
“The second stage is working with manufacturers. Today we are working with three of the 10 largest heavy equipment manufacturers in the world. The plan is to sell them our computer and have them integrate it into the machine assembly line.
“In the final stage, the goal is to work with the system integrator of the autonomous platform. There are a number of joint ventures today consisting of the manufacturer, the communications platform provider and the hardware provider – for example, companies such as Siemens and Bosch. The joint venture chooses to implement solutions from different providers within the autonomous platform; an information security provider, a lidar (light detection and ranging) provider, and we want to be the provider of the algorithm related to the image processing of the camera provider. Our goal is to become a company that sells a license to use advanced algorithms, but at the moment, our focus is to integrate and grow into the construction and mining industry.”
While the company is making progress in real-time alerts at construction and mining sites for heavy equipment, there is no doubt that there is a wide variety of additional applications that can be derived from the algorithms it has developed.
“Since our computer knows how to ‘talk’ to different types of cameras, we are already connecting it to a camera installed on a crane at the construction site in order to transmit alerts from an overhead view of possible dangers. We can even install the system on maritime cranes.
“We are currently operating pilot projects in the shipping industry with one of the largest ports in the world. Beyond that, we are also conducting pilots in the field of agriculture, to show relevance in parallel markets, but there are currently no plans to sell in these industries in commercial quantities – we remain focused.”
Connects to any camera
“There are two main trends in these areas that require attention: very low levels of safety and lack of efficiency,” explains Tzach. “Although these industries make up only about 10% of the world’s workforce, their contribution to accident statistics resulting in death is over 50%. This causes workers to stay away from these industries – especially the younger generation, who prefer to work in more advanced and technological industries.
“In addition, productivity per worker in these industries has not improved over the years, which is in contrast to the growing demand for new infrastructure and construction projects,” he adds, “In fact, according to McKinsey, every five days, more than a million people move from rural to urban areas, with an emphasis on Asian countries. This creates a massive demand for infrastructure, residential buildings, and the like.”
Initially, INTSITE focused on developing an AI-based system designed to make the crane, one of the most essential and dangerous tools on construction sites, autonomous. At the same time, it developed the ADAS systems. About a year ago, the company decided to focus solely on ADAS systems for heavy mechanical-engineering equipment, allowing the company to advance significantly in its business development.
According to INTSITE, the advantage of its system is that it connects to any analog camera and does not require the installation of expensive lidars and other sensors, which are a significant barrier to sales.
“Heavy vehicles are becoming more sophisticated and more autonomous,” explains Tzach. “In the world of autonomous vehicles, technology is classified from stage 0 to stage 5, where 5 is fully autonomous, and 0 provides alerts. Our company chose to focus on the first stage, meaning that we give real-time alerts and know how to give signals for control of the equipment, if there is a safety or operational need to slow down, stop, or change the direction of the machine.
“Our product is a ‘black box,’ containing a computer with powerful computational capabilities from NVIDIA. The computer can communicate with all types of cameras, which is the only sensor we use. We don’t use lidars, radars or any other expensive sensors, which is a significant advantage, as it significantly reduces the costs for our customers, as well as allows us to work with new machines that are already made with cameras, or with existing equipment on which we have already installed cameras.”
As previously mentioned, INTSITE takes advantage of the fact that cameras are already installed in close to half of new and existing heavy equipment, which can be connected to the computer. The black box on the other side connects to the operator’s screen, through which visual and audible alerts are given. In addition, a cellular module uploads the data to the cloud for monitoring by decision-makers. The software and algorithms add value in a number of areas, including safety, efficiency, preventive maintenance and reduction of fuel consumption, helping the environment.
“Regarding safety, the software provides the driver and those near the equipment with real-time alerts,” explains Tzach. “The software has over 30 different algorithms, which provide alerts in a variety of situations – for example, if the vehicle is going to overturn or run over someone, things that unfortunately do happen.“Beyond that, we are streamlining the work process by giving simple and intuitive explanations to the equipment operator while getting the job done.
“Work efficiency also leads to energy efficiency and a decrease in fuel consumption. These vehicles consume hundreds and thousands of liters of fuel a day. The financial savings are significant and, of course, benefit the environment.
“The last issue is preventive maintenance. By using video, we know how to detect, predict and alert before malfunctions develop that disable the equipment. We do all these things in real time, using only cameras.”The company offers several services in one product and can provide detailed reports to site administrators, thanks to the computer’s connectivity to the cloud.
“The processing is carried out on the computer in the equipment, but also in the cloud,” Tzach notes. “Everything that requires real-time alerts, such as safety hazards and damage to production capacity, is provided in real time, regardless of cellular reception, and everything that does not need immediate reporting undergoes processing in the cloud and is shown to management.
“The product complies with all relevant standards regarding information security and privacy – a very important issue since collecting information from cameras is very sensitive, especially in Europe.”
INTSITE owns two patents in image processing and artificial intelligence. The ability to process the image allows the computer to recognize and track objects; and through artificial intelligence, the software teaches itself to develop contexts, improves and becomes better and more efficient.“The basis of everything we do is data, which means any footage from the relevant areas in which we operate,” explains Tzach.
“This has been strategic for the company and for me from day one. We have installed cameras for no charge at various sites around the world so that we can collect the data and develop the algorithms. In order to build an algorithm that prevents a vehicle from overturning, I need to film scenarios where the equipment will overturn and teach the software when it will happen and when it won’t. Once it’s accurate enough in terms of accuracy percentage, I can release it to customers.
“Image processing technology has its limitations,” he notes. “In extreme weather conditions such as snow or heavy fog, there’s a chance the camera won’t see anything. But it doesn’t affect us as much, since we’re not dealing with autonomous vehicles now, and the equipment cannot be used in extreme weather conditions. In any case, we were able to collect data even in extreme weather conditions, so there is a certain ability to work in these conditions.”
The growing global need for ADAS systems for heavy mechanical-engineering equipment stems from inefficiency in the operation of mechanical-engineering equipment that becomes the bottleneck on the production floor – most often in mining or construction. This inefficiency stems mainly from the difficulty in monitoring these machines and the fact that such equipment is used in areas thousands of dunams in size.
“These traditional industries have been digitizing for the past two decades,” concludes Tzach, “and they need creative solutions like ours to meet their specific needs, at a very low and competitive cost.”
This article was written in cooperation with INTSITE.
Translated by Alan Rosenbaum.