Let’s go back to 28 June 2007. A day before the launch of the first iPhone. This event changed how the internet was experienced. Up to that point, we were stuck to desks and computers if we wanted to browse the web. Or at best, were using sub-optimal experiences using Blackberries and Palm Pilots! The graph below says it all. Looking at the growth of data traffic compared to voice, the paradigm shift is clear. In China, Xiaomi, built a multi billion dollar business by calling its brand Mobile Internet.
In short, what made phones smart was the core technical architecture that allowed access to the internet on the go.
Dumb vehicles versus smart mobility
Given the massive size of the transportation and auto industry, the innovations in this market are going to affect a wide range of industries and urban and suburban life.
Where the comparison in the mobile phone market was between dumb phones (feature phones) and smartphones, the comparison in the transportation industry is between dumb vehicles and smart mobility.
The trends making smart mobility possible are abbreviated as CASE:
We could argue that the first two are essential, intrinsic characteristics and the second two are extrinsic characteristics.
Autonomous Vehicles (AVs) needs to be connected in order to transfer data in real-time.
For AVs to be really safe and useful, autonomy, prediction and action all need to happen without any failure.
Shareability and electric trends make them more viable, without which the obstacles to innovation and adoption are simply too large.
Our focus here is on the autonomous part. Without sensors to provide awareness of the surroundings, vehicles are blind and need humans to operate them. This ties in with the 5 levels of autonomy announced by the Society for Automotive Engineers.
Once you equip a vehicle with sensors, you need to process the output of that sensor data to get actionable insight about the path while ensuring a safe and enjoyable riding experience. This is called perception engineering and is showed in the left middle box of the AV stack in the picture above.
Big obstacles to making vehicles smart are lack of data, engineering and operational expertise in the perception engineering and autonomous part of the system. This is still the case but the problem has moved on to known unknowns (corner and edge cases) which still act as the biggest barrier to adoption. The output of the perception part is needed for artificial intelligence (AI) systems to predict the right course of action and without the high quality data, these models will not be good enough.
In short, the smart component for vehicles is access to perception insight (sensor data+AI) in the same way mobile internet (universal connectivity+apps) made phones smart.