Will cognitive systems change connected vehicles as we are designing them today?
From today perspective the 5 disruptive technologies (sharing economy, cloud computing, digital supply chain, 3D printing, and Internet of things) will design a connected vehicle enabling new business models in the automotive industry. But cognitive computing is not listed yet.
In most connected vehicle cases, cognitive computing is in discussion either as
- In-vehicle concierge providing driver with natural language dialogue and advice on interacting with the vehicle for route planning, environment information, etc.
- Mechanic’s Advisor supporting drivers and mechanics in complex vehicle diagnosis and maintenance.
Can we enter with cognitive computing into safety areas of the vehicle? Safety and security is the most demand area of the connected vehicle market. Some passive cases are already known like early safety recall detection by analyzing data from NHTSA, warranty, call center, social media etc. Or reactive cases like automatic connection to emergency services in the event of a crash.
But most of these cases can be solved with traditional programmable systems and some deeper analytics capabilities. So what is the new thinking of cognitive computing from a connected vehicle perspective?
Without connectedness (vehicle-to-vehicle, vehicle-to-infrastructure, vehicle-to-cloud, etc.) a self-driving vehicle can’t be implemented in a regular, public traffic. A programmable system is deterministic and all rules are well defined in a captured context. Self-healing or self-learning is out of scope. A programmable system might be able to control the vehicle how to drive. But where to drive is much more complicated. Just a parking lot is marked different across countries. Also a huge knowledge base is needed to identify different objects involved in the traffic.
Yes, computing is getting more and more powerful. Think that today computers runs at about 5 GHz and consumes energy of 100 Watt per square centimeter. Compared to that, our brain runs at about 10 Hz and consumes energy of 10 milliwatt per square centimeter. In addition, the current computer architecture separates memory, computation and communication which is integrated in our brain. We need to think more about the required resources if we further design and develop the new ear of connected vehicles.