HERE Technologies

AI, Machine learning and self-driving cars

Recently, while working with high definition maps (HD Maps) and HAD Highly autonomous driving) concepts, noticed new approaches are emerging in our design, discovery, prototyping activities.

Designing experiences for AI and Robots will require completely new approaches, rarely we find new talent from Academic world trained with these approaches. For example, Engineers that design algorithms for AI, will need to design training modes for interaction design. That is, machines (robots, cars, etc) will need to have a great level of self-definition and adaptation, after shipping. Driver behavioural data is the normalising training input for AI engines, which can shorten the learning and baselining other alternative inputs.

On the other hand, Designers require to be working with data and passive behaviour as key input, and standard interaction design with preset logic flows no longer apply, instead Situational awareness cues need to be gathered and presented when they are relevant.

These are some of the recent tasks I am leading:

  • Leadership of engineering and design team delivering advanced positioning solutions for next generation location content (HD Live Map).
  • Strategy creation and development of concept designs for platform services.
  • Development of research agenda for situational awareness concepts assisting HAD modality, including computer vision and LIDAR for input in mixed mode rendering combining sensor fusion, HD map content and services such as traffic, IoT/Smart Cities or personal behavioural data feeds.
  • Adapting and prototyping rendering engine running on Nvidia embedded hardware and real-­time design demonstrator tools to serve use cases during highly autonomous mode.

Some thoughts on where are we heading, because may be obvious to some, as we see some signals on the horizon.

The intersection of Automation, Electrification and Smart cities/e-Cities can lead way not only to how we think about transportation, but also energy, or just our mental model of the definition of what a Car is.

  • Uber, not needing drivers, can achieve higher uptime use, only limited by the charge cycle downtime, the cost per km/mile can go significantly lower than with a driver, also the social impact may be considered in different terms. Nevertheless, we still need to think where these jobs will go… We have seen some of this, on airplane fleets.
  • Service providers, in order to achieve higher efficiency on their business would need to manufacture their own vehicles, or have strong partnerships (GM+Uber) as the TCO for the fleet operating will be governed by lower cost for fleet. Vehicles may need to be re-engineered to be maintained, if not fully recyclable, that is termination of the whole vehicle is not an option, but also how the cars are designed to be 24h in use (if you remove the battery swapping time or charge time).
  • Cars could generate energy too, not only consume, when cars can park themselves in solar-parks and plug themselves to the grid, once they are fully charged these can be contributing energy not only to other vehicles requiring charge (even in remote locations) but also to major industrial consumption centres. This will offset the need for nuclear power, fosil-based energy as many cars are parked most of the time.
  • Consequently, Parking space in city / urban context becomes less and less needed, as cars can park themselves, we find more space for human life, bicycles (or other human-carriers/pods), personal vehicles are also placed on centralised automated parking, for same reasons on not only charging efficiency, but also buffering energy and freeing space in urban centres.
  • For these, to happen is important for vehicle manufacturers to start looking at the vehicles as mobile servicing assets and best way tho to think about the car not as a product but as a unity that can buffer (charge, store, release) energy. The CSR agenda for these car manufacturers speak about sustainability, imagine they could eliminate altogether the energy they use, but also produce energy to other Energy utility consumers/industries.

Some closing thoughts on the business model side:

  • The vehicle (note I wont call it car anymore) could be provided for free, with transactional model maybe even fully free supported by content consumption in side cabin or even outside, or for a yearly subscription fee. The first model is already experimented by Google, with others Tesla, Amazon or Apple who may follow as they have no legacy and their culture may support these.
  • Yet, in support of traditional manufacturers point of view, we still see the need for the  Experience of Driving as something that won’t go away, as that is the core of certain product segments in todays’ leading brands. This does not impede the ability to for example Porsche to charge a yearly fee for a 911 Driving Experience, which may or may not include ownership as part of the Experience.