Chief Strategy Officer, Aileen Smith, discusses the benefits of embedding analytics at the ‘edge’ when it comes to critical functional safety applications

Since the early days, our industry has seen the pendulum swing powerfully back and forth in terms of the location of processing power: centralized mainframe computing, distributed desktops, centralized / regionalized cloud “compute and storage”, and now the latest trend towards the edge…

 

With analytics and big data driving so many applications and insights which help us run our lives, it’s increasingly important that we pay attention to the recommendations in this excellent article by Bernard Marr, and carefully plan where we process data, processing it at the most efficient place, and not unnecessarily spending money to move data around.

He gives many great examples where edge analytics can be very powerful, saving time and money by processing data and making decisions as close as is commercially viable to the source(s) of the data.

But for some applications, doing your analytics at the edge still isn’t good enough. In my view, the pendulum actually needs to swing even further – to what I call the “VERY edge”…

Imagine an autonomous vehicle travelling in an urban environment at 30 miles per hour. A software application at “the edge” which detects and reacts to anomalies – for example by applying the brakes – would be hard-pressed to detect and react in under 100 milliseconds. A few hundred milliseconds is probably more realistic. However, in only 100ms the vehicle will have travelled 1.3 meters. When you think about it … 1.3 meters is actually quite far! If the anomaly detected is a human in the street, 1.3 meters feels TOO far to have travelled before the brakes are applied, particularly when you consider the standard braking distance at that speed is over 12 meters.

Performing the analytics in hardware, at the ‘VERY edge’ could detect and initiate a response to the same anomaly in a fraction of the time … measured in microseconds rather than milliseconds. This could mean the difference between life and death.

UltraSoC’s unique technology offers this ‘VERY edge’ hardware solution, and at UltraSoC we’re currently exploring a number of such critical use-cases with our customers and prospects. Exciting times lie ahead as we work together to find better solutions for security and functional safety applications, working closer to the edge than is otherwise possible.

To learn about how to ensure safety and security from the ground up and the importance of embedded intelligence in safety and security applications, download UltraSoC’s white paper here.