A device with a .50 caliber machine gun on treads, with infrared sensors, programed to shoot at anything within range from 270 degrees to ninety degrees, to either quit firing if it is turned around, or keep track of where the front lines are if it is turned would be rather daunting of the bad guys in front of it. The guy running it, or everyone has a kill switch. Perhaps the good guys would have some identifying transmitter that the device would not fire at.
I confidently foresee new and interesting legal challenges as a result.
I’d confidently foresee rich pickings for those in the legal profession, but that isn’t much of a prediction.
If we’re going by Rule of Cool, I think the money could be better spent on R&D for Iron Man armor.
Yes, there’s always an upside! 
It seems like it would be easier and quieter to develop a kevlar vest for a mule.
I’ve been thinking about setups like this a little overnight, and it seems to me that any unattended killer robot we’d build right now would be too primitive to be useful. If you make the system only fire over a certain arc, then the enemy will be likely to eventually get behind that arc and capture it. Then they get a free .50, and free tech transfer. Aerial bots have this problem, but to a lesser extent. You have to shoot it down first, and a lot of the remains get damaged in the resulting crash.
If you used an ID transmitter, after the sensor is reverse engineered, it’s a location token telling the enemy where to shoot. That’s a weakness of the mule bot I linked to above, it’s following a transmitter on the soldier. I presume that the range of that transmitter is short, but a .50 can hit targets a mile away (accurate at about 1/3 of that). Any transmitter that is protecting you from one is advertising your location to everyone listening for a pretty long range. Alternately, they can just jam the frequency(ies), removing your protection from your own weapon.
So, I think that for the near term, armed robots are going to be very limited machines that will require a person minding them in order for them to be useful. If nothing else, a person will have to be around to tell it when and when not to unleash hell.
I’ve only dealt with horses and burros myself, but my wife’s great-grandfather had mules. Just getting them to pull a plow was difficult. If you left them unattended for a second, they’d drag everything back to the barn. I can’t imagine trying to get them to go anywhere when there were explosions going off.
Unlike the mule, the robot doesn’t crap everywhere and hopefully gets sick less often. Most importantly, it completely lacks a sense of self preservation. It’s only desire is to follow the transmitter.
Plus, the Army has lots of mechanics, and few muleskinners these days.
But, in the end, you’re right, that thing is damn noisy.
The Mule craps everywhere because it can find fuel while you’re on the move. I question how fuel efficient a mechanical four legged creature is over rough terrain, while stomping at the ground like a blind boxer with a grudge against the Mole Men. Animals can at least be a little fuel efficient because we move pretty fluidly and do a good job of moving or redirecting all energy in a single direction, beyond the minimum for getting up and over obstacles.
It probably is worth seeing just how far they can get this technology, even if they know that it will cap out below the level at which it becomes useful for practical purposes. But I suspect that the key result won’t be an electronic pack mule. Most of the technology will probably go into base-bound, externally powered robotics; forklifts that work on soft soils, power suits, artificial limbs, etc. Probably the current Robomule design is best suited for something like base defense. Figure that it can go indoors and up and down stairs. Give it some way to detect friend and foe, and it’s going to cause an awful problem for anyone trying to go up a staircase of the US embassy that it doesn’t like even if you don’t put guns on it. The racket might even be a bonus, from the standpoint of making it difficult for the enemy to operate.
It isn’t a fuel efficient design, either. I haven’t seen a walking robot I would call efficient, and that engine sounds like a two stroke.
However, you really can’t give the Mole Men an inch.
(snipped)
I think you’re generally correct, until this part. Solving the friend-or-foe problem is going to be hard, especially when you’re going to want it packaged in a mobile unit that can do something other than just ID friend or foe. Humans are far from perfect at it, and computers aren’t even close to humans.
I’ve been in the armed services. Parades and Passing In Reviews are a waste of time. There are better uses of enlisted men/women’s time than that.
What do you think is the aim of parades and passing in reviews?
What would be the point of having robots do them?
Well there’s a few easy techniques.
[ul]
[li]Everything in front of me = Enemy, Everything behind me = Friend[/li][li]Everything outside this perimeter = Enemy, Everything inside the perimeter = Friend[/li][li]Everything that my controller is shining a fancy-spectrum flashlight on = Enemy[/li][/ul]
Probably the first big solution would be to break a base down into regions and, similar to closing off bulkheads, you evacuate a region and mark it “enemy”, sending the beasts in to kick the ass of anything that goes into any of those zones.
Seconded.
Also all that “winning hearts and minds” stuff. Have ED-209 go break bread with the tribal leaders.
COM·MIS·ER·ATE
COM·MIS·ER·ATE
Yeah, I know Daleks aren’t technically robots
The first two aren’t really aren’t using any more discretion than a minefield does, and the third is using a human (I assume that’s what the controller is referring to) for identification.
Berserker - Fred Saberhagen.
Yes. Better granularity would be better, and I suspect possible, but I think we could both agree that these options are technically feasible, reliable, and useful. You could very quickly turn robomule into a defense device, using these techniques.
The point would be to show them off.
Well, I agree you could make it, and make it reliable. I’m just not sure it’s increased effectiveness and ease of deployment justifies the cost increase over a claymore mine. However, the worst thing about mines is that they’re cheap, and armies leave them behind. Make them expensive, and they won’t be leaving it behind.