Sensor Research
Sensor Summary Depth range Detection Time Sensitivity Power Requirements Weight Cost Notes/opinions
Nuclear Quadruple Resonance Detects specific chemical compounds in a manner similar to MRI. 7.6cm to 25.4 cm 0.5 to 3 minutes (stationary) ~0% false positive; 99% effective ~30 lbs $5,000 Metal mines shield internal explosives from detection; Cannot detect TNT in the presence of AM radio stations
Hyperspectral Imaging Takes a range of images that aren't in the human spectrum, creating images with signatures of objects that don't match the background. These objects can then be searched for by size and difference from background material. Surface 8hz Around 85% detection with 3-4 false alarms per image in a relatively harsh clutter environment 12VDC @ 5 amps 9.7kgs $85k- $500k
Water Jets Uses water jets to create a high pressure, but low force that goes into the ground. The sounds that come up from the holes are processed and it can be determined what the jet stream hit. 12” depends on length of of squirt, about a second per 6" 93% effective 220V/ 440v ~30 amps -> need an engine to run pump 50 lbs ~$1,500 for pump
Chemical Vapor Sensor Uses Amplifying Fluorescence Polymer (AFP) to detect trace levels of explosive substances. Presence of an explosive substance reduces intensity of fluorescence of the polymer, and a photodetector senses this light reduction. Extremely sensitive (possibility of high false positive rate?) $28k Can cover a broad area; not so good at individual detection
Neutron Methods Fire a bunch of neutrons at the soil and watch for radiation indicative of Nitrogen. Neutron Moderation is the only viably portably technique, but this technique will also detect hydrogen in soil moisture. 10-20cm several seconds per scan Theoretically high, but unstudied dedicated generator or highly radioactive material needs a dedicated SUV $100k No human portable solutions exist yet
EM Induction (metal detector) Detects the presence of metal with inductive coils 20cm constant detection High, but equally high false positive rate Usually 9V or 12V, <25W 0.5lbs-5lbs $30-$30,000 Standard demining procedures usually require investigation of all hits from a metal detector
Ground Penetrating Radar Well established technology for detecting buried objects. Lightweight, fast. a few centimeters in wet soil, dependent on soil moisture (20cm for anti-personal, 40 for anti-vehicle) false positive <25% ~12V; 30W ~5kg $23,500 needs around 5m separation from other sensors in environment to work optimally
X-Ray Backscatter Light and portable detection of mines at shallow depth ~10cm limit High detection time, though robot can use more powerful generator than human and thus reduce this very few products geared towards mine detection
Infrared Mines retain and release heat at different rates that their surroundings. Take thermal IR images and analyze thermal contrasts to locate mines. Used mainly to detect antitank mines on roads where environmental conditions are minimalized. Possibly useful to detecting hotspots in the morning before sending robot out. 10-15 cm 5 min to 1+ hours depending on time of day to analyze differences in rates of heat loss/gain, could also use single snapshots (much less accurate) 90% detection rate for homogeneous soil without vegetation 1-2 kg $1k-20k Effective only in certain hours/weather in homogeneous soil. Easier to detect mines if they were recently buried (within couple weeks). Requires highly sensitive (capable of detecting temperature variations of 0.1 degrees) thermal IR camera.
Acoustic and Seismic Systems Play signal on a loudspeaker. Watch surface for vibrations with displacement sensors. Very low false positive rate and high rates of detection with sophisticated signal analysis. Unaffected by weather/moisture. One mine diameter depth limit 125-1,000 secs to analyse 1m^2, though time can be reduced by N with N sensors 95-100% detection rate, No false positives with good signal processing 10-200 kg Technology is still mostly in the research phase, few/no vehicle mounted systems exist; Currently useful as a confirmation technique unless can build an array of displacement sensors, at which point might function as primary technique.

Other Methods

I left out several detection methods on due to lack of information or large size and/or energy requirements. Bio sensors were also left out due to unsuitability to short-term use by a robot (they require growing things or training animals). The complete list is:

  • Conductivity
  • Electro-magnetic induction
  • X-ray fluorescence
  • Nuclear Quadrupole Resonance
  • Gamma ray imaging
  • Electromagnetic radiography
  • Ground penetrating Radar
  • Laser/Lidar/Lips
  • TeraHertz imaging
  • X-ray backscatter
  • InfraRed
  • Bio sensors
  • Acoustic, seismic systems
  • Neutron methods
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License