Building Blocks for Developing Contextually-Aware Vehicles
TEAMS Design Trend Story Automotive III
The idea of a car that not only can autonomously drive from point A to point B, but also is intelligent enough to understand itself, it’s driver, and it’s environment and make decisions that benefit all parties involved is an enormous challenge. The race towards solving the complex web of factors involved has intensified greatly over the last few years, and yet no single solution has been fully developed and brought to market.Let’s take a look at the primary building blocks for developing Contextually-Aware Vehicles, and some of the greatest obstacles yet to be solved.
A. Data Collection
One of the largest challenges to solve in the beginning stages of creating a truly Contextually-Aware Vehicle is developing a comprehensive and adaptive data-collection system. Without reliable and thorough means of collecting information, lack of input and inaccurate data turns into more of a burden than a help to a driver / passenger. Which factors are most important to understand, and how can information be collected about them?
- Understanding The Vehicle: We all know about the change oil icon… the “low on gas” warning… the dreaded check-engine light… But what if cars were able to quantify and present more transparent diagnostic reports to it’s drivers? Why shouldn’t cars be able to predict and understand when something needs to be changed, and inform the driver with more in-depth and valuable reports? Almost no one would doubt the value of a vehicle which can analyze and diagnose its own status (other than maybe mechanics profiting from generic warning signals) … Yet creating tech and software that are able to measure and record the state of each object and system in a car is a gigantic task. Bringing each of these systems together onto one platform where they are even able to speak to each other is another task entirely. While great advances are being made in terms of onboard monitoring systems, creating a cohesive and ecosystem of instruments in charge of constantly analyzing the vehicle and it’s components will take quite some time.
- Understanding The Driver / Passenger: A huge topic being explored right now is how a vehicle is able to capture data from drivers and passengers. Capturing the digital identities of passengers and their preferences can lead to custom / tailored content and possible route suggestions. Evaluating a driver’s emotions will lead to a more tailored experience that empowers a driver to calm down when angry, or music suggested when sad. Understanding a driver’s physical state through advanced driver-facing cameras and health monitoring systems will prompt more responsive distraction warnings, or even emergency response warnings and manoeuvres. Improved Audio Input technology and Digital Assistants in the car will empower passengers with advanced interactions with services, as well as the ability to initiate commands (from making appointments, to finding parking spots, to making purchases). Developing the hardware and software to understand who is in the car, and what their motivations are will lead to the creation of driver behavior models and powerful driving experiences.
- Understanding The Environment: The last essential data-points needed for a car to grow contextually aware have to do with everything else going on in the world around it. Through advanced traffic information systems (V2I Technology), pre-configured and crowdsourced mapping data, satellite imagery, current and upcoming weather measurements, news events, a host of sensors constantly monitoring and predicting the behavior of other road users… The list of external factors relevant to a driving machine in terms of data capture is tremendous. Until a vehicle is able to sense, capture the complex relationships between all of these factors, reliably and securely, the fully Contextually Aware Auto will still take quite some time.
Not surprisingly therefore, the web of interweaving data points required to enable a truly Contextually Aware Vehicle is a gigantic endeavor, one already known by auto manufacturers and tech / software providers… It’s not that intensive work isn’t being done in all of these fields. Rather, the difficulty today is that car companies are trying to provide individual solutions for individual problems instead of creating cohesive systems for capturing, syncing and understanding all of the data points.
B. Data Processing
Once one has solved the fundamental areas of collecting such a gigantic stream of vehicle, driver and environmental data, how does one even begin to sort through and understand the ongoing massive data flood? Some organizations are predicting that cars alone will generate around 4000 GB a day… To put this into perspective, that’s more data captured and created in a single day than a human consumes or generates in a month.
“Vehicles will generate and consume roughly 40 terabytes of data for every eight hours of driving” — Brian Krzanich (Intel CEO)
The ability to store, process, analyze and draw valuable insights from the various data generate by cars is one of the fundamental challenges to be solved. We’re not just talking about huge storage units… this involves a host of technologies still largely in development. From Multipoint Control Units and Internal Networks channelling information from every part of the vehicle, to intricate computational powerhouses (much like the brain) charged with sorting, storing and equipped with machine learning algorithms to filter the information generated, to 5G Networks facilitating the rapid flow of data from a car to the cloud.
The ecosystem of data capture, collection and processing technologies involved are called Automotive Processing Platforms, and several automotive suppliers (such as NXP & Qualcomm) as well as hard/software companies (ie. Google / Intel) are developing rich and powerful systems to handle to sheer volume created by cars in the future. Organizations will prosper or perish based on the way that they are able to process data generated by vehicles, and how they convert that information into meaningful contextual information for cars.
C. Decision Making
Data is useless on it’s own… and the final imperative building block for developing truly Contextually Aware Vehicles will be the ability to discern and act based on information gathered and processed. Since autonomous vehicles are still in the early stages, much of the information gathered will not be able to be understood directly by the car itself.
Rather, the task now is processing and presenting the information in ways that are understandable and helpful to a driver or passenger. Many technologies are being developed, from advanced Infotainment platforms which present contextual / curated content to passengers, E-Cockpits & Dashboards displaying more intelligent diagnostic statistics to drivers, and Heads up Display technology which allows drivers to keep their eyes on the road while still receiving custom notifications near their windshield area.
In the near future though, further and further driving decisions will be handled by the onboard processing units inside vehicles themselves. The need for low-latency, reliable rapid decision making and data transmission between various systems inside the vehicle will be essential before vehicles are trusted to fully operate based on contextual decision making. The culmination of identifying relevant contextual information on a timely basis, and often making split-second decisions in an array of complex road situations will form the basis of Automotive Platforms.
Unfortunately, we are here still greatly in the beginning stages of empowering vehicles to make decisions as well as humans. The Media is rife with story (Waymo), after story (Google), after horror story (Uber) of halfway autonomous vehicles not far enough in being able to identify and handle complex road situations, often ending in tragic injuries. But even now, one might be able to make the argument that vehicles are turning out a lower rate of accidents based on systems in their current state than the multitude of accidents caused by human error around the world today. Only if, and when, the combinative vehicle technologies can sort through the flood of relevant data packets to react for the safety and benefit of all road users will drivers and passengers be able to fully trust that their vehicle and the decisions it makes on their behalf.
Clearly, although the challenges ahead are known to various degrees and greatly in progress, exciting advances and pieces of this interesting puzzle are slowly coming together. The sum of these building blocks, and the path to truly autonomous-driving vehicles in the future will be paved by solving the challenges of creating a Contextually Aware Vehicle.
If you’d like to read more about concepts regarding Contextually Aware Vehicles, here are a few more articles with some good thoughts on the topic:
What the heck is context aware technology?
If you read my previous post about context aware cars, you know my preference in automobiles and digital snitches. You…
Driverless cars: mapping the trouble ahead
When a self-driving car looks at the world, there are many things it sees. It has radars that measure distance to the…
Autonomous driving: digital measurement of the world
Even in the future, sensors will not be able to supply every detail about their surroundings; it is always possible…
5 big things still standing between us and a glorious self-driving car future
It's fun to ponder a future filled with self-driving cars, a world with breezy commutes where robot navigators have…
- The Sentient Car: Context-Aware Automotive Telematics
- Design of Context-aware Systems for vehicles using complex system paradigms
This is part of an ongoing series we at TEAMS Design are producing to discuss future trends we are researching in the automotive industry.