Envisioning A Data-Enabled, $2 Trillion Automotive Ecosystem Before Autonomy Is Real

The automotive ecosystem is an almost $2T marketplace which consists of a large number of integrated markets.  Beyond the automotive OEMs, these include rental companies, auto financing, auto insurance, gas stations (energy), media (radio, billboard in particular), maintenance services, public sector infrastructure, and even emergency services.  

Figure 1:   SAE Edge Report:  Transportation Ecosystem 

Autonomous capability has been touted as the disruptive change agent by the media and investors alike. However, autonomy has proven to be very difficult at a technological level.  As “Measurable Safety, The Missing Ingredient To Demonstrating ADAS Value” discusses, even ADAS, the simpler poor cousin to advanced mobility systems, is not ready for prime time.  Is there a technological layer which underlies both of these capabilities, that may actually be the first to provide real value to customers ?   

In an automotive capability stack, below the perception systems and machine learning systems sit the core sensor, communication, and control systems.  Today, nearly every new car has these three systems embedded as a part of its core functionality with associated cellular communication systems back to the OEM HQ. Note, these communication systems do not have to be nearly as reliable as those required for safety critical active driving operation.

With these basic atomic capabilities, the automobile can participate more fully in higher level use-models with help of higher levels of software applications. Further, the infrastructure can be similarly upgraded with sensor, communication, and control capabilities. The combination allows for a data and software system which can much more fully optimize all sorts of interesting use-models.

What might be some interesting value-adding use-models ? 

At the individual level, friction points include: 

1. Parking:  In dense urban settings, finding an open parking spot can be challenging. In this new world, smart infrastructure can maintain a map of open parking spots. Provisioning, reservation, and payment can be done on automobile interface. No need for parking meters, coins, and perhaps even physical signage. Alerts on parking time expiration can directed to the driver’s cell phone.
2. Maintenance:  With a smart car, automobile sensor systems can perform predictive maintenance. If maintenance is required, the owner can allow access to the maintenance vendor. Vendors can take automobile and leave a replacement which is automatically provisioned to the owner.
3. Fueling:  Why is it necessary to go to a gas station ?  Why not have the gas station come to you with a subscription for refueling ?  Not only would this be more convenient, but likely saves aggregate fuel if done at large parking lots.   
4. Delivery:  Today, delivery is to static addresses, with location services, one can enable delivery to your car truck. It just requires remote permission to trunk and location services.  
5. Add-on Devices:  Today,  tolling agencies, insurance companies, and local security have to build, distribute, and enable add-on after-market devices for enabling their services. With a software based approach, all of this can be performed by simple enablement by the driver.
6. Media:  Billboards and radio are especially effective because of their localization. However, remembering the content of a carefully constructed advertisement while traveling 60 mph is not so easy. Billboards and Radio can communicate the critical aspects of location and offer to the panel of the automobile. In this manner, the driver just selects the advertisement and it can be added to the mapping system.

Whatever the minor concerns of an individual owner, the problems are amplified for fleet operators and enabling software services are even more valuable for them.  Finally, transportation management agencies can leverage this data to build a deeper understanding of the environment (congestion, weather, emergencies, etc) and take corrective actions. Currently, public agencies have to build replicative investments to provide these basic services.   

Today, the automotive OEMs are certainly generating the data (in various formats) and they have access to cellular networks to transmit the data, but is someone pulling together the data and software middleware to enable the optimization of these broader processes ?

Is it even possible to align a massive ecosystem such as automotive to enable these services ?

A startup in Israel called Otonomo is taking on the challenge. Otonomo has built a platform which claims to have a strong architecture of privacy (opt-in) and federated secure access. With this platform, they can take data from automobiles, and make it available to application developers for constructing larger flows in an organized manner. 

“We have over 24 million cars and agreements with 12 OEM, and over a 100 commercial agreements for application services,” said Ben Volkow, founder and CEO of Otonomo.  Pulling all of this together is not easy.  According to Ben, key challenges include:

1. Normalization: Aligning data meaning across the various sources.
2. Cleansing:   Removing nonsensical information caused by anomalies. 
3. Enrichment:  Fusing car and contextual data to build a cohesive picture.
4. Trip Aggregation:  Build higher level information for applications.
5. Triggers and Access:  Limited control access to the automobile.

Meanwhile, all of this must be done with a strict view towards consent management and security. 

Overall, as was described in “What Can Amazon And Hilton Can Teach Uber About Transportation,” transportation has been missing the benefits of software controlled assets. Cloud vendors such as Amazon AWS
AMZN
have built a large pool of middle-ware which can enable the management of a large number of assets with very high efficiency and scalability.  Otonamo seems to be moving transportation in this direction and the benefit for customers could be substantial.