Nomad

What Nomad is

Nomad is a four-wheeled robot developed at Carnegie Mellon University's Robotics Institute with funding from NASA ^[1]. The university describes it as a planetary rover prototype ^[2]: a machine built to work out, on Earth, how a rover could drive across unfamiliar ground and decide for itself which rocks were worth studying. It is roughly the size of a Volkswagen Beetle and weighs 725 kg ^[1].

Machine and instruments

In its field trials Nomad moved on studded tyres at up to 50 cm/s, with power supplied by a petrol-driven generator ^[1]. Navigation rested on three kinds of sensor: a pair of stereo cameras, a laser rangefinder and GPS ^[1]. The stereo pair gives the robot two views of the same scene, and the difference between them yields depth; the rangefinder measures distance directly; GPS tells it where it is. For the science itself Nomad carried a high-resolution camera, a spectrometer and a metal detector ^[1]. A spectrometer spreads incoming light into its component wavelengths, and the resulting pattern can be read as a clue to what a surface is made of; a metal detector responds to metal-rich material. The robot also has an arm, which lets it place scientific instruments precisely on the rocks it has chosen to study ^[3].

Searching the ice

Nomad's search pattern is deliberately simple: it moves up and down across a search area in the way a person operates a lawn mower ^[1]. Within that pattern it looks for rocks that stand out by their dark colour against the white ice ^[1]. Those dark stones are the targets. The robot's job in Antarctica was to find meteorites lying on the ice, and a member of the ANSMET field team would go out and collect any meteorite Nomad located ^[3].

Doing it alone

The Antarctic expedition was Nomad's fourth overall, but the first on which it was expected to move and analyse specimens completely on its own; earlier outings in Antarctica and in Chile's Atacama Desert had been system tests ^[1]. That shift, from a robot driven by its operators to one that picks its own targets and analyses them, was the point of the exercise. Whittaker, quoted in a Carnegie Mellon field release, said the breakthrough technologies are robotic classification and search, and that both must succeed for autonomous meteorite discovery ^[3].

Where it sits in the development of rovers

As a prototype, Nomad's value lay less in how many meteorites it found than in the questions it forced into the open: can a rover cover a region systematically, recognise a target by its appearance, place instruments accurately on that target, and produce an analysis with no person deciding each step ^[1][3]? The instruments it carried are the ordinary kit of field geology. The hard part is giving the machine the judgement that usually belongs to the geologist, and letting it act on that judgement far from anyone who could take over.