Boundary RiderThere may be a couple of basic flaws in the design...

Boundary Rider
There may be a couple of basic flaws in the design of the Autonomous Surveillance Vehicle which need to be addressed.

One is that the information from sensors needs to be analysed at the edge, not sent off via the cloud to a distant CPU. Partly a matter of security, by avoiding hacking, and partly to reduce latency. A machine used in mines must be self-contained, as communications could become very tenuous. All of this indicates the need for a neuromorphic chip, such as Brainchip’s Akida, to carry out edge AI.

The second flaw is the single mast. Nature does sensors very well and usually has two (at least), not only for redundancy but also for stereoscopic analysis. What happens to our Boundary Rider if some lag puts a bag made of foil-faced building insulation over the mast. Is it blinded? Can it still transmit? If one heads toward the goal of edge AI mirroring as much as possible the mutli-modal zero latency ideal, one could end up with two seats on the boundary rider, in which two guards sit (brilliant neiromorphics), as the machine goes round the perimeter at 25K/hour (30 minute circuit). The humans are there to act quickly if a breach is found, and can handle the visual analysis, and the machine can handle the X-ray and electronic surveillance, and to provide visual recording. And a 30 minute spin, on a random schedule, would not be too arduous or boring.

Any thoughts, as we seem to have been at this problem for some time, without any sales.?