We’ve done a few posts on port-side robotics. I’m all in on the concept because the scope for industrial grade IoT is massive. But, once we flip the script across to sidewalk surfing things begin to kerb dive. The problem is an evolving, constricted, contested environment. More than that, there is a hackneyed narrative shaping a theater which is not a reality, but a simulacra. A badly written marketing pitch that wants to be read as destiny.
A delivery robot is not much more than a cargo focused remote control car. We’ve had those since the mid 60’s. I was mowing down lego figurines with a metal bodied RC Ferrari in the last century; a savant or a robotics pioneer? A media push to get people to gape and goggle at that seems absurd, so some sleight of hand is underway. We’re not talking smart devices, but sharp practice. Technocracy is very comfortable making humans behave like machines, while insisting machines are becoming human.
“Pie in the Sky”
People seem to be slow in catching the vibe. It is simple. In reality that teenage guy on the Glovo bike is paid a little over minimum wage to transport food to your doorstep (hot) using his cognitive faculties and fast twitch reactions across a problem ridden domain. He is the evolution of hunter-gatherer and the pony express rider, not R2-D2. The key business (Read: Profit motive) skill is in transferring as much risk as possible to the rider and the consumer.
Automation succeeds where human risk is maximal, space is controlled, liability is clear and social negotiation is minimized. That is why our high value sectors are ports, mines and warehouses, not sidewalks. This is also why ag robots with NASA stickers demo well, but slide off in practice, like those monster machines you see in agricultural shows. The concept is again marketing oriented; to demonstrate overwhelming “power” to puny human farmers. Unfortunately, the “gigwork Glovo” example works even better in food production than it does in food delivery.
Many, if not most, of the commercial applications for delivery robotics involve “human-in the loop” (HITL) design. The concept is not pushed to media, because it ruins narrative flow, but it is where the real “gravy” is. Symbiosis is practical. It allows a spectrum of options, instead of a linear preconfiguration. “If your system only works when reality behaves, your system sucks…”so they fudge reality, using terms like “supervision” instead of control to skirt liabilities and unfavorable optics.
But, how are they “driving” remotely?
Not with an Xbox controller in a basement (anymore). Operators typically get multi-camera stitched views, LIDAR overlays, depth cues, bird’s-eye maps, sensor and condition indicators, latency-aware controls, predictive smoothing (inputs buffered and corrected and a comfy seat. Think air traffic control meets drone piloting, not RC toys.
To wrap this delivery; the trick is to keep HITL intact, not to hide it. It is a value proposition, not a hidden extra; delivery drone driver has a bit of a ring to it. Car automation is cruising a smoother track than sidewalk crawlers, because the core constraints have been in place for 100 years. A stray dog may still run out into traffic or a deer through your windshield (depending on location) but the terms of engagement are clear, liability is understood. There are “rules of the road.” In that scenario your robot cannot “be rude” to my Shih Tzu.
