The Aftermarket Analyst, August 2017
August 28, 2017
Show all

The Automotive Analyst, February 2018




Peering into the Autonomous Future

Hopefully you’ve gotten used to writing “2018” by now, or typing it at least. We thought it might be fun to take a moment and reflect on some automotive topics, trends and technologies we think will be important this year but that we may or may not quite understand and try to make sense of them. No headline-grabbing here, just some clear, honest straight talk, in no particular order.

Autonomous cars will be here any day now! Or sooner, or possibly later…but wait, they’re already here!

Does anybody pay attention to predictions anymore? There are some great roundups here, here and here. The variance of the predictors raises more than a few questions, including, but not limited to: should there be a penalty for overestimating? Who’s being sincere and who’s head faking? And what exactly are they predicting?

Ahh, what does “autonomous” mean? There’s a cornucopia of “autonomous” features available today, including automatic braking, lane-keeping, adaptive cruise control, and auto-parking to name a few that automate certain tasks but you can’t check out and take a nap yet. For that you need a “self-driving” vehicle, which is what it sounds like: a vehicle that drives itself once a destination has been input and can make decisions on its own. A car with autonomous features isn’t necessarily a self-driving car, but a self-driving car necessarily has autonomous features, although those features don’t necessarily make it self-driving since there are other things that make it drive itself. Clear as mud, right? That might explain why surveys regularly find people are mistrustful of autonomous vehicles. When people who aren’t plugged into the industry hear “autonomous” or “self-driving,” they don’t think in Levels 0 to 5. More likely, they think pods without pedals or steering wheels (driverless vehicles!), or some other manifestation of the thing which scares them, like zombie cars prowling the highways searching for unsuspecting pedestrians. Surveys also show people don’t always trust or understand the autonomous technology in their current vehicles. (Is that correlation or causation?) The only sure thing is the marketing department has its work cut out for itself.

My advice to somebody who thinks somebody else is using the incorrect term: be kind. It doesn’t really matter all that much because in the end everybody’s going to get where they’re going anyway.

Is more power the key to Level 5? (Computing power that is)

Ford CEO Jim Hackett recently said his “business is computing.”[1] Another Ford exec, Sherif Marakby, this one in charge of autonomous vehicles and electrification, said that computing power is “going to be one of the fundamental enablers to getting autonomous vehicles to operate in just about every environment.”[2]

If your vehicle has a voice recognition system and it’s frustrated you to the point you never use it, it’s because of a lack of computing power. And for autonomous vehicles to operate in all environments, not just in well-mapped or geofenced areas, the vehicle will need even more computing power. Even with the advent of 5G, a vehicle won’t transmit ALL 4 terabytes of data it generates each day to the cloud. It will need to be able to analyze the data onboard and then transmit only what’s necessary.[3]

Nvidia estimates autonomous vehicles will need 50 to 100 times more computing power[4], just announced a computing platform codenamed Pegasus that it says can be used for Level 5 automated driving. Pegasus can perform 320 trillion operations per second, which is 10 times faster than its predecessor.[5]

If you want to give your brain a workout, read up on Quantum computing. The computers we know and love do one thing at a time, but a quantum computer can do many things at the same time. For example, a conventional computer can be A or B at a given moment but a quantum computer can be A and B, a difference that compounds exponentially and allows them to complete massively complex calculations in speeds previously unfathomable. In late 2015, scientists at Google and NASA, working with a D-Wave quantum computer, solved a complex computing problem 100 million times quicker than a conventional computer.[6] You may never personally use a quantum computer, but the company that’s routing you and another thousand autonomous or driverless vehicles away from that traffic jam on I-95 will.

The importance of 5G

The next generation of wireless communication, 5G, promises speeds up to 10Gbps, or about 10,000 Mbps versus your home internets’ paltry 25-100 Mbps. It will be the ultimate in cord-cutting. 5G’s most impressive party trick might be its low latency, or the time from when the message is sent to when it’s received. In June, Nokia, the Finnish consumer electronics company, announced it lowered network latency to less than 2 milliseconds using its 4.9G technology.[7] Current LTE networks are capable of around 25 milliseconds and 5G is expected to go even lower. Lower latency is an important step in delivering real-time services to autonomous vehicles, or for Vehicle-to-Vehicle or Vehicle-to-Infrastructure communication. It could even allow for remote operation of vehicles.

It cannot be stressed enough how important 5G will be to the development of autonomous vehicle technology. It can facilitate the kind of real-time communication required to make self-driving work and managing the data derived from autonomous and self-driving vehicles.

Somewhere in time and space

In addition to computing power and communication skills, Level 4 and 5 self-driving vehicles will need to know a lot of things. Like where it is, where it’s going, how fast it’s going, how fast it can go, what are the potential obstacles, are those obstacles mobile or fixed in place, what are the road conditions, are they hazardous, is it cold enough to freeze, will that pothole damage the wheel, or is it a puddle, is it safe to swerve, is the vehicle working properly, when will it need to refuel or recharge, how many occupants are there, are they all sentient, are there any quirky local driving habits to consider, and so on and so forth. It’s going to require a combination of cameras, Radar and Lidar.

You generally need 4 to 6 cameras to make realistic 3-D images. These images are essential to identifying things like traffic signs and signals and lane markings.[8] Radar, short for Radio Detection and Ranging, is already used by driver assistance systems such as blind spot monitor, collision avoidance, and distance control. It looks for things that may cross the vehicle’s path. Lidar, short for Light Detection and Ranging (and is sometimes spelled LiDAR, but then why isn’t Radar “RaDAR”?), is used to measure distance to objects, essentially it’s making a map of the vehicles’ surroundings. Not everybody, *cough* Tesla *cough*, thinks Lidar is necessary. It’s true that Lidar is currently expensive, bulky and stymied by bad weather, but there are some very smart companies working very hard to solve those issues and the fruits of that labor is already evident. It can also be argued fusing all three types of sensing adds a necessary layer of redundancy that increases safety and accuracy.

Infrared and ultrasonic sensors are useful for short-range tasks like identifying lane markings and objects. GPS and inertial navigation systems can be used to aid in localization. High Definition maps add even more context to the sensor data and are continuously improved by drivers sending their data to the cloud. Accurate maps also happen to know what’s on the other side of that hill, while the sensors only know there’s a hill coming up.

What’s Next?

What technology excites you? Is it 3-D printing and the notion of “printing” your own auto parts? Or maybe the developments in battery technology? Or augmented reality that projects navigation, alerts and other data onto the windshield? The possibilities, and opportunities, are boundless.


Notes

[1] “CEO Jim Hackett says Ford is more than a mobility company: ‘My business is computing'”

[2] “Ford’s Marakby says computing power key to self-driving success”

[3] “Connected Cars Bring New Business Models and New Disruption”

[4] “Nvidia unveils ‘world’s first’ AI computer chip for fully autonomous vehicles” and “NVIDIA Announces World’s First AI Computer to Make Robotaxis a Reality”

[5] “Nvidia says its new supercomputer will enable the highest level of automated driving”

[6] “Google and NASA Say Their Quantum Computer Finally Works”

[7] “SK Telecom reduces latency to 2 milliseconds for 5G” and “Nokia shows how 4.9G technology will reduce network latency on the path to 5G at #MWCS17”

[8] “Three Sensor Types Drive Autonomous Vehicles”




Recently Announced Automotive Aftermarket Transactions

PrecedentsFeb18




Selected Public Company Capitalization and Operating FiguresPublic CompsFeb18