The autonomous automobile seems to come and go as a subject in the media as does the autonomous truck for shipping. This trendy topic has a lot of concepts that affect engineering and design. With over fifteen relevant references written on this subject within the last year alone, there was no shortage of information. The focus is on three major issues, the amount of data required, government intervention, and the players in the industry. Each of these subjects would make for a separate blog discussion. For this blog, we shall stick to the data related issues.
Although analog engineering plays a large part in enabling autonomous vehicles, the major argument is around the large amount of data needed for accurate decision making. Combine this with the ever-increasing amount of data being generated by advancing automobile intelligence and the result is a communications overload problem that makes for a great engineering discussion.
Communication has gone from basic analog signal transmission to a bit data rate per second with various solutions boasting about rates of Gb per second. Meanwhile upstream, there is more and more data being generated4.
Just one autonomous car will use 4,000 Gb of data/day
Self-driving cars will soon create significantly more data than people—3 billion people’s worth of data, according to Intel
Even more intense data generation figures are predicted. “It was reported by Ford that connected car sensors generate 25 Gigabytes of data per hour, and then by the WSJ that a typical autonomous vehicle generates 4 Terabytes of data in 90 minutes, and then by Intel of 45 Terabits per hour. All are massive numbers, but why so different? It points to a bandwidth problem. The raw data is beyond any auto OEM’s ability to manage, even in 5G, and so the amounts reported could be raw data or some prioritized data, that facilitate core vehicle diagnostics and operating behavior improvements8.”
It remains to be seen how Terabits of information will be transmitted on current technology that is limited to Gb/sec. The majority of this data has to be communicated internally within the vehicle as well as externally. Furthermore, the vehicle will require location signals such as those from Global Positioning System (GPS) transmission as well as sensors in smart roads.
Within the vehicle, bus overload is already becoming an issue. Communication is going beyond the typical CAN bus to Ethernet levels16. The increase in the number of sensors in a vehicle is going to complicate this problem even further as each sensor generates more data.
When communicating externally to the vehicle, the limitations are based on the lack of bandwidth in the overcrowded frequency spectrum17. There is no room to add to the current allocation of frequencies. Those frequencies that are dedicated to data transmission are seeing exponential increases in the amount of data that is being transmitted. The data is increasing faster than the hardware can handle it.
Vehicle locating is a huge issue with the autonomous vehicle. Radar has limitations especially over distance. Thus, sound-based RADAR technology is giving way to light-based LIDAR technology. Lidar, or Light Detection and Ranging, operates on a similar premise as radar. However, LIDAR takes things a step further by sending out a prearranged series of light pulses at regular intervals9. The speed of light offers relief in location technology local to the vehicle however communication to the data centers will still face the problem of an overcrowded frequency spectrum.
The scary part about all this data is that processing priority is left to the software. “Theoretically, as software companies, with massive real-time data, they can choose to optimize data processing between vehicle operating behavior and the environment beyond the vehicle8.” Let’s just hope that two autonomous vehicles prioritize the fact that they are headed towards each other before alerting the driver to a low tire pressure due to cold weather.
Regardless of what investors think, software will not be the solution to this problem of “data guzzling” as some refer to it. New hardware with faster communication capability is needed. Data transmission needs to be increased by several magnitudes if the autonomous systems are going to work.
Automotive ignitions systems have seen many transitions over the years. Historically, the designs have matured from a magneto to today’s coil-over-spark plug designs. The progression follows the emergence of solid state electronics as well as the phasing out of mechanical components in favor of electrical components.