The cars of today are significantly more advanced than the ones made even 10 or 20 years ago. Most of these advancements have been made under the hood, without the users really knowing or understanding what changed. One such adjustment that needs to happen is a move from physical conductors to wireless connections for data transfer in cars. ADAS cars have many barriers to hurdle, and weight and space requirements are one of them. An explosion of sensors and the move to centralized processing in cars only exacerbates this problem. A move from conductors to wireless will reduce the weight of smart cars and leave more space inside the car for other components.
The Bandwidth Problem
Everyone knows that widespread adoption of automated vehicles will represent an Internet of Things (IoT) explosion, and will also lead to a new era of “Big Data.” The thing is, all that data has to come from somewhere. It will come primarily from all the sensors involved in operating ADAS features, and there are a lot of them. Many of these sensors will need to transmit huge amounts of raw data to be processed, hence the need for new high bandwidth connections in cars.
Whether you’re designing a traditional car with ADAS features or a fully autonomous vehicle they’ll both have one thing in common, sensors. One recent self-driving robot failure highlighted the need for multi sensor fusion in autonomous systems. New cars will be packing everything from LIDAR and/or radar arrays to passive visual and ultrasonic sensors. Intel estimates that connected cars will need to process up to 1 GB of data per second. New cars will need communication systems that can handle all the information coming in from a huge variety of sensors.
Some sensors by themselves will need high bandwidth connections. Take passive visual for example. Your car may have several HD cameras constantly feeding raw high-quality images or video to your electronic control unit (ECU). So not only will the system as a whole need good connections, individual sensors may also need them as well.
ADAS enabled cars will use lots of sensors which will produce large amounts of data.
The Weight Problem
If ADAS enabled cars need better connections, why not just use more copper? The problem is that with all of those distributed sensors you may have to add significantly more wires. Wiring already takes up a huge amount of weight and space on current cars, so increasingly that load isn’t really an option.
In the past, some manufacturers used distributed microcontrollers to process data. Now, though, as processors grow more powerful and AI is being introduced to cars, centralized processing is becoming commonplace. Imagine having to add more wires from each far flung sensor to a central ECU. The car will be chock full of copper.
In fact, cars are already crammed full of conductors. A premium sedan can be packed with 12,000 ft of wires weighing up to 110 pounds. Remember that there are other things that have to fit inside cars aside from sensors and wires. All those miles of cable take up space that could be used for other features. Some of those wires have to stay, as we’re still not ready to power things like motors or actuators wirelessly. A few people have proposed solutions to reduce the size and weight of the other cabling. One such solution is using Ethernet. Its proponents say it could reduce weight by up to 30%. However, I think with wireless systems we could get that percentage much higher.
We don’t want the insides of our cars looking like this.
The Wireless Solution
Two problems, bandwidth, and weight, but one solution: wireless. A company called Keyssa recently announced that they’re making a chip that shows us how a wireless system in a car could work. This system would have the bandwidth to transmit data from multiple sensors and would greatly reduce the number of wires in cars.
Keyssa’s integrated circuit (IC) is a close proximity wireless connector that claims data transmission rates of up to 6 Gbps. That kind of bandwidth would certainly be enough to send raw video footage or a point cloud from a sensor to a central module. They also say it’s low power, which is another concern in ADAS cars. These chips have general applications in the IoT space, but I think the technology they showcase is particularly useful for cars. The only problem is that Keyssa’s specific product seems to be made to work in very close proximities, like within inches. The primary reason for that appears to be so that the user wouldn’t have to pair two devices using the technology, it’s “pairing” is done with distance. I doubt, though, that pairing would be a problem in a car, and the wireless modules could be hard coded to communicate with the processor.
This kind of scheme would obviously solve the weight problem. No wires means no weight. By using a wireless system like this you could eliminate all electrical wiring that wasn’t delivering power to components. That’s a large reduction.
We’ve always known the future would be wireless, so let’s go ahead and update the communication systems in our next gen cars. The way things are currently both bandwidth and weight/space could become big problems. Wireless systems are an elegant solution to both problems.
After you’ve decided on your in-car communication network you’ll need to program it. TASKING produces software that’s made specifically to help you design programs for ADAS enabled vehicles. Products like their standalone debuggers and static analysis tools are meant to help you do your best work as efficiently as possible.
Have more questions about wireless systems? Call an expert at TASKING.