Valtna had previously worked for 20 years at the university, but by that time she had gone to Milrem, one of the most successful future makers of our defense industry, as the head of autonomy development. They had an ambitious plan to make their robots move intelligently in the terrain, i.e. in the grass. In fact, it is a very difficult task, because there are no suitable sensors for this yet.
A team of four from LightCode Photonics came up with the idea of creating a camera that sees in three dimensions. This would help solve many problems in the field and overcome bottlenecks. Existing lidars are expensive, break easily, and are computationally intensive.
"Lidars are actually the first snow that has started to melt. We say that we make a 3D camera with dynamic pixels. So that machines can move in the world meant for people," Valtna says.
Bad news for those waiting for self-driving cars
In the new company, it was decided to take a customer-oriented direction. As early as possible, they wanted to know what problem someone needed to solve. In this way, technology development can be better managed. It soon became clear that many robotics companies do not have suitable sensors. Bogdanov reminds us that 3D display technology initially gained wider coverage in connection with self-driving cars. It was like an early warning that something like this was coming.
"We quickly realized that the problem there is much more complicated, and it doesn't make sense to solve it in the first place," states Bogdanov. However, they saw that service robots, especially those that operate indoors between four walls, must become much smarter. Sadly the truth is that the world is quite far from self-driving cars. The technology may already be very good, but it either lags behind performance or the price is too high. "This prevents robotics from triumphing," says Bogdanov. Therefore, it has been difficult to give robots, for example in factories, a large number of tasks that are routine or dangerous for humans. As LightCode's CCO Terje Eichelmann concludes: “Robots must save money and increase productivity. They must act quickly, painlessly, and safely. The last place is the key because robots have to move in the same environment as humans without injuring themselves and others.”
When a technology company gets one of Bolt's first investors as its investor, it must be something special. It is still a matter of people belonging to the first 150 rich people in Estonia. Valtna recalls that when it was possible to meet people again after the first corona wave, an investor, who does not want to be named in the media columns, came to their laboratory for the first time and was very surprised. Why?
Simply put, LightCode Photonics takes a single pixel and increases its resolution. In this way, all 128 pixels can be given additional depth information. This is their uniqueness. This investor currently has ventured into 3D vision, mobility, drones, vehicles, and virtual reality in the portfolio.
"He said it was the holy grail," recalls Valtna. "There has been a lot of support from it, but of course, we have to pull the chestnuts out of the fire ourselves," she adds.
Getting to the lntel accelerator in Munich has also given them confidence. There are strong mentors and a nice energy. The ecosystem there has left a deep impression on Valtna. In Germany, it seems that without a doctorate in physics, chemistry, or engineering you are not allowed to such events. Many of the founders talk about their Tesla background there. This seems to be the standard in old Europe.
"We are the same as the German tops. If you are in the right place, your uniqueness is the norm. It's awesome to experience," exclaims Valtna, and also praises the depth and knowledge of the investors. So many times they have been asked very high-level questions at the table. "Then we feel that we are doing the right thing."
Reaching an important milestone
According to Valtna, LightCode Photonics will soon reach a very big and important milestone. Their device becomes demo-ready or enters the next phase of testing. It's kind of been a long journey, but they've been building the device for two and a half years. In January 2021, when it moved into its premises, the technology fits into a large metal box the size of a suitcase. It weighed a lot, the total price might have been 100,000 euros, and it only worked in a dark room. Now, two years later, it also works in daylight and is 180 times quieter. The price has come down by an order of magnitude.
Bogdanov adds that the step has been really big. The power has increased significantly in this short time. "The device that will be used now is particularly unique. It's entirely our design," he says. We are already moving from the demo phase to the production phase at a faster pace.
Several problems with vehicles
Where can I find their cameras? Especially from logistics, warehouses, and last-mile solutions. In their case, the whole world is now wondering how to do it well and safely with robots. Robots must save money and increase productivity. They must act quickly, painlessly, and safely. The main keyword is the last one because robots have to move in the same environment as humans without injuring themselves and others.
Let's get back to self-driving cars. Bogdanov says that there has been an exaggerated enthusiasm for them. He does not claim that such vehicles will not come, but the time gap is misjudged. This is not just a matter of sensor technology. Software, algorithms, and social agreement are also a problem.
Software, algorithms, and social agreement are also a problem.
"The social aspect also comes into play. How much are we willing to hand over to robots?" he asks. Currently, it makes more business sense to use sensors indoors because it is a more controlled environment. Bogdanov sees the robotization of factories as the next big step, a revolution.
The last mile solutions, the best known of which are Starship and Clevon, are also very much looking forward to new cameras. In general, if the task is well-defined, it is possible to perform it with a robot. In this way, the price of the service or product for the end consumer can also be kept down. Micromobility and drones may also be in the sights of LightCode Photonics in the future.
Gives the robot more confidence
The logic of Tesla and other big players has long been that a person has two eyes that can be replaced by two cameras. In marketing, however, the fact that we also have another special organ, the brain, was left out. It has not been able to be generated yet. Until robots have brains like humans, they have to compensate with different data. 3D data helps this world significantly. This gives the robot more confidence to make decisions.
"Currently, the circuit is such that if the robot cannot decide by itself, then a person brings his brain into play to make a decision. The robot needs better data and certainty. The camera technology is not comparable to the eye. The human eye is an incredibly sharp organ," says Bogdanov.
Milrem and the defense industry would also be a good outlet for them, and the company is their partner. According to Valtna, it is good to test solutions with them, because the situation in which these military robots operate is again very difficult. Cooperation is also carried out with the space technology side to test the camera at high altitudes as well.
The intersection as an air traffic control center
One of the uses is also smart city solutions to plan new smart intersections. In addition to self-driving cars, we also get to those vehicles that have been talked about in Estonia too: the rental car is driven to you from a distance, and you don't have to go to the car. Bogdanov says it is possible and the process is ongoing. The state and the developers discuss the matter with each other. This car is not running. He would be controlled from a distance. Gradually, however, life will also lead us to self-driving cars. However, the co-founders of the company do not see a sudden change - neither technological nor social - in this.
Safety is such a critical issue that the head of Waymo's lidar department, formerly known as Google's self-driving car branch, became interested in the Estonian company's activities. Namely, they have edge cases where they need very good visibility to see far. For example, it is necessary to find out whether the car is going on or off the highway.
"Perhaps the infrastructure itself will lead us instead. It is not practical to put so many sensors on cars because the speeds are high. Maybe in the future, every intersection will be like a technology hub in an air traffic control center."
The university has a very big role
LightCode Photonics currently employs 17 people, nine of whom have PhDs. However, the founders do not emphasize it too much, because of course it is not needed everywhere and, for example, this degree is not given in business management. "As a human, you are still the same biological machine," says Heli Valtna.
Jan Bogdanov adds that if we talk about deep technology, this knowledge is not the kind that can be obtained from the Internet. They don't get googled. Companies often grow out of universities because the original hardware is expensive and is concentrated in universities.
"LightCode wouldn't exist if there weren't people with PhDs," he is convinced.
Why are there more and more such companies in Estonia? Because even here in this tiny corner of Europe, we have the opportunity to finance them. Valtna considers the ecosystem very important. He agrees with what he heard recently, that there is no envy in the Estonian technology sector. Here we support each other. In addition, there have been investors with experience and experience who have opportunities, and interest in new directions.
"We have a fresh way of thinking, resources, and interest," sums up Valtna.
The company currently has four patent applications pending and plans to file more.
/ The article is translated from: https://arileht.delfi.ee/artikkel/120140662/eesti-ettevote-arendab-uht-tehnoloogiamaailma-puha-graali, originally published by Tanel Saarmann at Delfi Ärileht /