Startups are more knowledgeable about autonomous driving

It seems that all players in the autonomous driving field have agreed that the actual implementation of autonomous driving will be between 2020 and 2025. However, not all players in the autonomous driving field can clearly explain how the autonomous driving technology will be implemented. Let's follow the automotive electronics editor to learn more about the relevant content.

Startups are more knowledgeable about autonomous driving

For startups, it is really unrealistic to rely solely on investment to survive until the 2020s when autonomous driving technology is fully implemented. Therefore, compared with traditional automakers with deep pockets, startups often have a clearer idea of ​​how to implement autonomous driving.

Two months ago, I interviewed Wu Gansha, CEO of UISEE Technology. Wu Gansha resigned from Intel and founded UISEE Technology in October 2015. He has been leading this startup for nearly a year and a half.

Startups are more knowledgeable about autonomous driving

In an interview in February, Wu Gansha said that he believes that the huge opportunities in the autonomous driving market are in areas that traditional large companies are unwilling to do or are difficult to do. In the short term, the core of UISEE's autonomous driving solution will still be low-speed autonomous driving vehicles in low-complexity road environments such as parks.

Last week, UISEE Technology held another media communication meeting. In addition to discussing the autonomous driving market space that has almost been talked about too much, it also discussed the most practical autonomous driving implementation plan.

Wu Gansha listed six trillion-dollar markets for the so-called market space for autonomous driving, namely the automobile market, travel market, logistics market, commercial real estate, energy market and the social benefits brought by autonomous driving. There is nothing wrong with what he said. Each of these six areas contains huge commercial value and can be related to autonomous driving.

Looking further, the development of society and history is actually linked to the development of transportation to a large extent. So from this perspective, as the current trend of future travel, autonomous driving obviously has a lot of potential commercial value.

But what needs to be done to fully implement autonomous driving? From a fundamental level, the entire transportation system needs to be changed.

This is not to deliberately criticize the current manufacturers of self-driving vehicles. In fact, if you think rationally, today's traffic system is actually tailor-made for human drivers. Traffic lights and lane lines are designed for human drivers' color recognition, car horns are designed for human ears, and speed limits and driving regulations are designed based on the safety range of human drivers.

Startups are more knowledgeable about autonomous driving

Let's take a look at what self-driving cars are doing nowadays. They use laser radar, cameras, and millimeter-wave radar to identify the surrounding environment of the vehicle, and transmit the environmental data to the on-board computer, which calculates and outputs driving decisions and controls the vehicle's movement. Among perception, decision-making, and control, the perception part can be said to be the least efficient, because self-driving cars need sensors to simulate human senses to read environmental information of the traffic system.

So what is a truly reasonable and efficient scenario for using autonomous driving? With the Internet of Vehicles, each car knows the location, speed and vehicle status of each other. Traffic signals are replaced by electrical signals that can be easily received by cars. Lanes are also closed as much as possible to avoid unnecessary human interference and reduce the incidence of accidents.

In other words, to achieve full-scenario autonomous driving is definitely not something that a traditional car manufacturer or any startup company can do. What is needed is the improvement of the entire system. So for current autonomous driving manufacturers, how should autonomous driving be done? In fact, the idea is very simple: to build a small-scale, low-complexity traffic system and autonomous driving in low-speed controllable scenarios.

At the communication meeting, the autonomous driving implementation solutions presented by UISEE Technology are actually a good direction. One is the autonomous driving shuttle bus in the airport parking lot, and the other is the autonomous driving shuttle bus project in the commercial real estate parking lot.

Startups are more knowledgeable about autonomous driving

In essence, both solutions are low-speed autonomous driving vehicle projects for low-complexity road systems within the park, but due to different environments, slightly different technical solutions are required. However, they are basically the same autonomous driving implementation solutions.

The difficulty of the autonomous shuttle bus at the Baiyun Airport parking lot lies in the stability and the impact of the humid environment in the south on the unmanned vehicle. The 4-seater unmanned vehicle has no steering wheel, accelerator or brake. Passengers get on the car with their luggage and just press a button, and the car will start automatically and go to the destination along the optimized route.

Startups are more knowledgeable about autonomous driving

The shuttle bus in the indoor basement parking lot is much more difficult. First of all, the underground garage cannot use GPS map to plan the route. In addition, the lighting environment in the underground garage also increases interference for camera recognition. However, the user's usage is basically the same as the airport shuttle bus. Just get on it and the self-driving car will take the user to the destination.

From these two cases, we can actually see that as an autonomous driving startup, in addition to its current focus on low-speed electric vehicles, the autonomous driving solutions that UISEE can provide are still quite flexible. When the environment cannot be changed on a large scale, making customized autonomous driving vehicles is still a good idea.

In an interview two months ago, Wu Gansha actually said, "As for the solution of the autonomous driving kit, UISEE will not adhere to a certain technology or solution. Instead, it will make precise adjustments based on user needs, acceptable budgets and usage scenarios." He also gave an example:

"In fact, we have always taken a very pragmatic attitude to develop our technology. The so-called pragmatism means demand-driven. In the future, I think QR codes are not necessary. Because you have a high-precision map covering the whole of China, you may not need a QR code.

The reasons for doing this now are: 1. In a residential area, we first test it. If the GPS and inertial navigation work very well, our work is done. However, if we find that the combination does not work very well and we find that there are large errors every time we walk a certain distance, we will post some QR codes on the roadside to help correct the errors. If the problem is solved, then it is OK.

Startups are more knowledgeable about autonomous driving

Or if I drive my car in an underground parking lot, there is no GPS at all. I cannot afford high-end inertial navigation. In this case, we may need to use laser radar or visual technology to solve the problem in a different way.

So we are very flexible. We have an algorithm library. Now the problem is that when a customer comes with a demand, we try it with different configurations.”

Of course, from the current perspective, whether it is the self-driving shuttle bus at Baiyun Airport or the underground garage, it is still more of a demonstration than a practical significance. The current driverless car is still in a demo stage, which means that the utilization efficiency of driverless cars is still very low. The number of vehicles that can be deployed in a scenario is not large, and each car still requires a specific safety officer to supervise it, etc.

But this still provides a direction for the development of driverless technology and time for the technology to develop. After all, people's patience is often limited. If autonomous driving makes people see no possibility of implementation, then the market capital will ebb as social enthusiasm decreases, and it will be difficult for the technology to continue to grow.

The above is about the road of autonomous driving in automotive electronics! Startups have a clearer introduction to this road. If you want to know more relevant information, please pay more attention to eeworld. eeworld Electronic Engineering will provide you with more complete, detailed and updated information.