Type to SearchView Tags

Edge Insights for Superior Autonomous Vehicle Experience
Ved Parkash Pati Business Manager, Digital & Analytics | May 7, 2020
177 Views

We are living in a hyperconnected world. Research shows that there will be a massive 20 billion connected devices in the world by 2020. This number is set to catapult to 75 billion by 2025.

Like many other industries, the automotive industry is also getting increasingly connected. Gartner predicts that by this year, i.e. 2020, an average connected vehicle will generate more than 280 petabytes of data on an annual basis, with 4 terabytes of data being generated in a day at the very least. Their research also states that around 470 million connected vehicles will be deployed by 2025.

Such data explosion through connected devices opens up rich avenues to glean and act on insights and make the autonomous vehicle experience of the future more data-driven, user-centric, and highly transformational.

And while, most of the automated vehicles today are either between Level 0 and Level 1, in order for autonomous vehicles to truly attain the highest level of automation and operational efficiencies, one emerging trend that verily promises to be the vital substratum is “edge computing”.

Edge Computing

What is Edge Computing?

Gartner defines edge computing as solutions that facilitate data processing at or near the source of data generation. The solutions include a robot, an industrial boiler, a smart meter, a thermostat, or a vehicle. This reduces the need to transfer the data back and forth from the cloud.

Edge computing is growing at a CAGR of 33% and is expected to have a market size of $16 billion by 2025.

Consequently, edge computing addresses the many limitations presented by cloud with respect to IoT, namely:

  • Low Latency as edge nodes will be present and no cloud round-trip will be required for processing data
  • Reduced Bandwidth as a local network of edge nodes doesn’t need to send all data to the internet
  • Increased Reliability as edge devices will be able to analyze data and provide decisions themselves
  • Improved Security as critical and sensitive information can be protected by local encryption
  • Cost Efficiency as initial studies show edge computing could lead to substantial cost savings over cloud

Courtesy the above pointers, edge computing is growing at a CAGR of 33% and is expected to have a market size of $16 billion by 2025.

Edge Computing in the Autonomous Vehicles Industry

The need for local processing and advanced data analysis make autonomous vehicles a good use case for edge intelligence.

By moving compute and processing closer to the source of data, it is possible with edge intelligence to glean insights in real time and help the vehicle with the “next best action” it should take.

Autonomous vehicles are different from other distributed and connected devices as they need “instant information”.  Thus, by having sensors in the camera, brakes, wheels, monitor and other areas, we can facilitate real-time collection, processing, and relay of information that would make the autonomous vehicle self-reliant, safe, predictive, user-centric, and ultimately, truly autonomous.

Here are a few use cases from the autonomous vehicle landscape where edge computing promises to play a defining role in the future:

Use Case 1- Air Conditioning Intelligence

We normally ignore car AC maintenance. And then our sweat pores open up and we understand that the car AC is in a condition of disrepair. We rush to our dealer who diagnoses the condition, paints a scarier picture, and charges a bomb which hurts our pocket.

In an edge world, the OEM (Original Equipment Manufacturer) can embed an edge computing device on the AC compressor connector. The tiny device with an ingrained analytics- and machine learning (ML)-driven model measures the pressure of the valves and cycles of the compressor connector, learns about the AC performance, and activates an SMS to the customer saying, “Hi. Your AC is performing at 70% efficiency and needs immediate attention”.

This is the most optimal time to go to a dealer with the intelligence that the AC is not as damaged (and hence wouldn’t warrant a pocket-burning fix). Compare this to today’s solution and see how edge can really make a difference to the car experience in the future.

Use Case 2- Battery Monitoring and Predictive Maintenance

The battery of the car should give the best throughput at all times. Therefore, continuous monitoring and predictive maintenance of the battery is important. The battery health is reliant on several factors like acceleration, traffic conditions, charging cycles (in case of electric vehicles), and others.

Edge devices can perform a real-time inspection of key battery parameters to perform predictive maintenance alerts that can inform the vehicle owner in case of any deviation. They can aggregate this data and monitor the key battery parameters. By leveraging edge computing, auto OEMs and network providers can deploy predictive maintenance and make a direct influence on the customer experience, especially in the electric vehicle segment.

Use Case 3- Hyper-customized Infotainment

Over the last decade, In-vehicle infotainment has taken a quantum leap through sleek human-machine interfaces, augmented reality, mobility apps, personalization, and other methodologies.

Edge computing can take the user experience to the next level by understanding applications and interfaces the user is using, making real-time optimization in terms of color, font, design, and resolution depending on whether it is touch interface, voice recognition, or a visual medium.

ML algorithms can continuously gather, process, and gain insights from the available data. These ML models can be hosted on the edge device to analyze the user behavior data and use this processed data to improve in-vehicle user experience.

Use Case 4- Tire Intelligence

Tire health is dependent on factors such as ambient temperature, mileage, tire inflation, braking, and acceleration, among others. If these factors are tracked in real time by the edge computing sensor tracker and by using ML intelligence, if a data model and a tire health score are created, then the moment the tired wheels give out a score below the acceptable threshold, a notification can be sent out to the driver and the wheels could be diagnosed at the next pitstop.

An email can also be automatically triggered to the tire brand who, in turn, can call up the customer and enquire for any further support.

Use Case 5- Smart Traffic Management

Imagine a traffic intersection which has four to five heavily used roads branching across each other. The autonomous vehicle cannot control the long waiting times at the traffic junction. But, in the future, if the road intersection has an edge device deployed to which all vehicles can communicate while coming towards the intersection, the edge device can aggregate the data from nearby vehicles and also suggest the most optimal routes to the vehicles in advance.

 This can help manage traffic in the most efficient way and have the roads unclogged at all times.

Use Case 6- Vehicle Security

In an edge and IoT world, every connected component in the vehicle could act as an additional layer of security authentication. So be it front lights, rear lights, cameras, and doors, each component could have a different password that could be configured as “mandatory” for the car to be activated and be able to run.

Also, a model where more than three consecutive wrong passwords send an SMS/email to the car owner can make the vehicle extremely safe.

Conclusion

The edge computing industry is at a tipping point and several large automobile manufacturers are busy creating and testing prototypes of their autonomous vehicle variants. As both the nascent industries mature, we will see a beautiful intersection of these two trends in the next five to seven years which shall pave the way for a superior autonomous vehicle experience.