Unmanned realization is inseparable from laser radar

Probe domestic switch needle KG-300K needle head diameter is 3.0mm normally open switch needle

Photocoupler

The auto industry always drums up some new words to torture us, thanks to it, now we know that driverless cars are "coming soon." Some manufacturers predict that autonomous vehicles will enter the commercial market, the earliest time is 2021. However, the price of new technology is very high, and most ordinary people cannot enjoy it for a period of time. Let's take a look at the related content with the car electronics editor.

Lidar (ALDAR) is so expensive that the so-called laser radar is an imaging system that allows driverless car software to see the road surface. High-end systems may approach $100,000 and low-ends will cost less than $10,000. Emerging technologies have never been cheap, but historically, manufacturers have often found ways to produce components at lower cost. However, driverless cars leave them too short, which means that if manufacturers want to launch a driverless car, they either accept high costs or delay production.

Unmanned realization is inseparable from laser radar

That's right, car manufacturers can choose cheap hardware to lower costs, but it's not worth the risk. Why? Because car autopilots put a huge burden on manufacturers, even mid-range lidar systems are as safe as gambling. According to the MIT Science and Technology Review, low-end lidar systems are useless when cars are traveling at highway speeds.

MIT uses two sets of data from the Velodyne system. The set is HDL-64E, which costs $80,000, and the set is Puck, as long as $8,000. The HDL-64E uses 64 laser beams to draw images for the surrounding environment. The resolution is very high, covering 120 meters at the farthest, and Puck has only 16 spots. The image will be distorted very quickly, and the farthest coverage distance is about 100 meters. If the car is driving at low speed, both systems are very good. If you drive at normal driving speed, Puck will not work.

The problem is that even the HDL-64E system is just meeting the minimum requirements for freeway travel. When the speed reaches 70 miles per hour, the car advances 31.3 meters per second. If an emergency occurs, the car needs a distance of up to 60 meters to stop completely. Every time you exceed one meter, the computer has more time to make decisions. It is entirely necessary for the car to be farther away. Ideally, a car using a lidar system would only provide an effective image 200 meters away, which is safe when driving at highway speeds.

One technology that can solve this problem is the solid state system, but solid state systems are much less expensive than traditional lidar systems. Quanergy claims that it has built a $250 component for the so-called S3 car. It is a pity that the system's fidelity is low, only when the car is driving at ultra-low speed (less than 10 miles per hour), otherwise it is useless. Velodyne is also developing cheap solid-state components, but it promises that solid-state systems have no way to replace 360-degree laser systems.

Luminar CEO Austin Russell explains that the sensors they use try to avoid solid-state hardware. Why? The main reason is that although the laser system is much more expensive, it has a lot of images to make driving safer. The sharper the image, the better. He also said: "About an object, if you only have a few points, even if your machine learning technology is amazing, it is no good, because you don't know what it is. If you can only see things 30 meters away, when the car is When the speed of the highway is running, it is a one second distance."

Graeme Smith, head of the University of Oxford's driverless program, believes that the laser radar industry may seek a balance between data quality and price, which can cause differentiation and allow time for high-speed driverless cars to drive on the highway. difference. He said: "Compared with high-speed driverless cars, the price of the low-speed version may drop faster. If you want the laser to cover a range of more than 250 meters, use a better laser. If you only drive in a low-speed environment. It can cover only 15 meters, so you can use cheaper sensors."

Still, installing low-quality lidars for cars will increase costs. Industry researchers and automakers believe that unmanned hardware will increase the final price of a car by $8,000-$10,000, but the actual cost may be higher. When you add all the hardware in your existing test platform, the final price is simply an astronomical number. Tesla said it wants to drop the radar system to $8,000, which looks beautiful, but it also installs thousands of dollars of computer equipment, cameras, and inertial measurement units (which can come in handy when GPS is not available). Many people worry that the images drawn by Tesla Radar are not clear enough.

In addition, if companies use high-definition lidar systems, they should switch to radar or shut down the system in fog. In an interview with Axios recently, Luminar's Russell said that if manufacturers want to retail a driverless car, they must do their utmost to ensure safety, and also reduce the "critical failure rate" to ensure that the image is extremely clear. Russell said that the driverless industry has placed too much emphasis on reducing costs, and now developers need to concentrate on enhancing the security of technology to make it unbreakable.

Russell believes that the cost of perfecting the technology may range from $300,000 to $400,000, which is what the fleet owner is willing to pay because driving sharing is a profitable business.

The above is about the realization of automotive electronics - the realization of unmanned driving is inseparable from the laser radar. If you want to know more information, please pay more attention, electronic engineering will provide you with more complete, more detailed and updated information. .