GT Sensor Arrays

Kenneth Witt:


In GURPS Vehicles, second Edition there are rules for building single sensors in any size from hand-held scanners up to massive multi-ton starship sensors, but there are no rules for combining the output from multiple collection points (individual sensors) to produce a more sensitive composite device with better resolution and a longer effective detection range than the individual components. There are many real world examples of such devices in use today. One of the best known is the Very Large Array (VLA) radio telescope. The VLA is composed of 27 separate radio telescopes spread across several square miles of desert in Arizona. The output from each telescope is digitally processed and combined together to produce an enhanced image of whatever astronomical radio source they are scanning. This process is known as aperture synthesis. Note that this is not the same process as synthetic aperture radar (SAR) uses to produce an image. A single radio telescope of sensitivity equal to the VLA and capable of producing comparable images would have cost far more to build than the array. In a similar application, I recently read an article about a low budget SETI (Search for Extraterrestrial Intelligence) project that is using an array of several hundred three-meter diameter satellite TV dishes to gather faint radio signals. Again the array of small dishes was much less expensive to build than a single large receiver of equivalent sensitivity.

Another example where multiple detectors out-perform a single sensor is in the towed sonar array used by most modern nuclear submarines and many surface warships. The signals from the many hydrophones along the tow cable are combined by a computer to produce a superior sonar image of all the sound producing activities around the vessel. The modern towed arrays can detect the presence of other ships at distances greater than World War II radar equipment. And because of the distance separating the individual sensors, the exact range to the target can be accurately triangulated by the processing software. Determining an accurate range to target is a very difficult task with a single passive sensor, but is fairly simple to do with some basic geometric calculations if you have directional data from at least two widely separated point sources. In GURPS Traveller space combat, the range and bearing information from a passive sensor net would be as accurate as active sensor data and have the same targeting advantages as active sensors.

One more trick that you can do if you have an array of sensors is called Long Baseline Interferometry (LBI). This can turn a sensor array with widely separated elements into a very sensitive motion detector by combining the outputs from multiple sensors and looking for the interference patterns at various bands across the electromagnetic spectrum caused by small frequency shifts in the energy radiated by or reflected from moving objects. Again this is an application where multiple small sensors can perform much better than a single large sensor of equal cost. The usefulness of this technique for detecting spaceships at extreme ranges should be very apparent to any Traveller player. If our real world technology continues to advance in this area, then in the near future we will have telescopic systems using this technique able to directly observe planets orbiting nearby stars.

The last advantage that widely separated sensor elements provide is the ability (by comparing the data from multiple sensors) to isolate and filter out local noise and extraneous signals from the final image data. In SETI research, which involves searching for very faint and distant signals, this capability to detect and eliminate local interference is absolutely essential. In GURPS game terms, the ability to filter out noise coming from one, or a few sensors, would make the sensor net as a whole much more resistant to jamming and spoofing than a single sensor.

Using currently available technology, this aperture synthesis process has been extended to optical frequencies as well with very impressive results. There are several technical issues which make aperture synthesis in the infrared and visible light frequencies difficult to perform with modern day equipment. The relative position of the individual sensors must be known with great precision (typically to within about half the wave length of the radiation frequency you are imaging), the individual data packets must be time-stamped with great accuracy, the communication between the sensors and the central processor requires a wide bandwidth connection, and the computer imaging process itself requires a very large number of computation operations. But, given the technological advancements in sensors, laser rangefinders, communications, and computers in the Traveller Universe, these problems become relatively trivial to solve.

Because of the obvious utility of sensor arrays composed of many small sensors and the lack of standard rules for the same, I have developed the following house rules for the building arrays of passive sensors (also called sensor nets) in my campaign.

Building a Passive Sensor Net


Step 1: Select a base scan rating and get the total weight, volume and cost from the PESA Sensor Table.

Step 2: Decide how many sub-units to divide the sensor into and select the corresponding software.

Step 3: Use VE2 to design a vehicle for the sub-units with appropriate propulsion and communications.

Step 4: Designate one or more processing/imaging/net control sites and the data link transfer hierarchy.

Step 5: Deploy the sensor network.


Step 1: Select a base Scan Rating and get the total weight, volume and cost from the following PESA Sensor Table.

Note that although TL10 figures are provided, they could be determined by simply doubling the corresponding TL11 values. TL9 figures can likewise be determined by doubling the corresponding TL10 values.

Table 1: PESA Sensors
Basic Scan
Nominal Range
In Space (mi)
TL 11+ TL 10
Mass Volume Cost Mass Volume Cost
(sTons) (dTons) (MCr) (sTons) (dTons) (MCr)
29 1,000 0.10 0.125 0.01 0.2 0.25 0.02 0.4
30 1,500 0.15 0.1625 0.013 0.26 0.325 0.026 0.52
31 2,000 0.20 0.175 0.014 0.28 0.35 0.028 0.56
32 3,000 0.30 0.2 0.016 0.32 0.4 0.032 0.64
33 4,500 0.45 0.2375 0.019 0.38 0.475 0.038 0.76
34 7,000 0.70 0.3 0.024 0.48 0.6 0.048 0.96
35 10,000 1.0 0.375 0.03 0.6 0.75 0.06 1.2
36 15,000 1.5 0.5 0.04 0.8 1 0.08 1.6
37 20,000 2.0 0.625 0.05 1 1.25 0.1 2
38 30,000 3.0 0.875 0.07 1.4 1.75 0.14 2.8
39 45,000 4.5 1.25 0.1 2 2.5 0.2 4
40 70,000 7.0 1.75 0.14 2.8 3.5 0.28 5.6
41 100,000 10 2.5 0.2 4 5 0.4 8
42 150,000 15 3.75 0.3 6 7.5 0.6 12
43 200,000 20 5 0.4 8 10 0.8 16
44 300,000 30 7.5 0.6 12 15 1.2 24
45 450,000 45 11.25 0.9 18 22.5 1.8 36
46 700,000 70 17.5 1.4 28 35 2.8 56
47 1,000,000 100 25 2 40 50 4 80
48 1,500,000 150 37.5 3 60 75 6 120
49 2,000,000 200 50 4 80 100 8 160
50 3,000,000 300 75 6 120 150 12 240
51 4,500,000 450 112.5 9 180 225 18 360
52 7,000,000 700 175 14 280 350 28 560
53 10,000,000 1,000 250 20 400 500 40 800
54 15,000,000 1,500 375 30 600 750 60 1,200
55 20,000,000 2,000 500 40 800 1,000 80 1,600
56 30,000,000 3,000 750 60 1200 1,500 120 2,400
57 45,000,000 4,500 1,125 90 1800 2,250 180 3,600
58 70,000,000 7,000 1,750 140 2800 3,500 280 5,600
59 100,000,000 10,000 2,500 200 4000 5,000 400 8,000
60 150,000,000 15,000 3,750 300 6000 7,500 600 12,000
61 200,000,000 20,000 5,000 400 8000 10,000 800 16,000
62 300,000,000 30,000 7,500 600 12,000 15,000 1,200 24,000
63 450,000,000 45,000 11,250 900 18,000 22,500 1,800 36,000
64 700,000,000 70,000 17,500 1,400 28,000 35,000 2,800 56,000
65 1,000,000,000 100,000 25,000 2,000 40,000 50,000 4,000 80,000

Step 2: Decide how many sub-units to divide the sensor into and select the corresponding software from the table below.

The software is required to integrate and refine the raw data from all the sub-units in the sensor net and display the enhanced images. Note the scan bonus and software complexity and cost from the following table. The more complex software can be used with smaller sensor nets at a reduced scan bonus. The software cost is for Tech Level 10 or higher. (Double costs at TL9 and double again at TL8 or lower.) Check the sensor table above for the revised nominal range.

Table 2: Sensor Nets
Number of
Signal Processing Software
Complexity Cost (MCr)
2-9 +1 5 0.008
10-99 +2 6 0.016
100+ +3 7 0.032

Step 3: Use GURPS Vehicles to design a sensor drone vehicle for the sub-units.

Design Recommendations:

  1. Use a small robotic brain (VE2, p61) and a robotic structure. (VE2, p18)
  2. Use NPU or RTG generators. Advantages: small size, no moving parts. (VE2, p86)
  3. Include at least three laser communicators for flexibility and redundancy. (VE2, p48)
  4. Civilian sensor drones should have radio transponders.
  5. Military sensor drones should have stealth and emissions cloaking. (VE2, p91-92)
  6. For "stationary" sensor nets, such as ones built to scan the space around star ports, thrusters providing 0.1G or less are more than adequate for station keeping. If the sensor drone has to accompany a ship, the capability to match or exceed to acceleration of the ship would be helpful.
  7. The vehicle design system in GURPS Traveller: Ground Forces could be used to simplify the design of a sensor drone with some small loss in flexibility.

Essential software will include Electronic Operations (sensor), Electronic Operations (communication), Pilot (small craft), Computer Navigation, Data Link, other?

Step 4: Designate one or more processing/imaging/net control sites and the data link transfer hierarchy.

Usually this is just a matter of installing the Signal Processing Software at a central site (either a starport or ship-board computer) and setting up a system to determine which drones relay data back to other drones and which ones talk to the central site. Usually only the closest two or three drones will talk directly to the central site. Drones that are further away use intermediate drones to relay their data to the central processing site. Of course, the central site could be a dedicated structure or ship designed just for this function. The communication system should be fault tolerant and provide for many alternate paths for the data packets to follow if one or more drones go offline. In essence, the sensor net should be designed as a wide area communications network. It is also desirable for the central site to have facilities to repair and refurbish the sensor drones.

Step 5: Deploy the sensor network.

Drones should normally maintain a separation of at least a few hundred miles. Wider separation between the sensor elements will actually improve the capability to perform LBI motion detection, but for simplicity in gaming, I have ignored this effect. Drones could be built in the same shape and volume of standard space interceptor missiles to simplify deployment from warships. Permanent sensor nets should be built with about 5% more drones than necessary to allow for down time due to annual maintenance and repairs.

Ad Hoc Sensor Nets

If you donít have any dedicated sensor drones, but you do have multiple ships with PESA sensors and copies of Data Link and the Signal Processing Software, then you can build an Ad Hoc sensor net. One (or more) of the ships is designated as a central image processor and runs the processing program. The other ship(s) use data links to transfer their raw sensor data to the central processor, and the squadron gains the benefits of a sensor netís enhanced detection capabilities. Because of the required precision in positional data between the sensor platforms, this is something normally done only in non-combat conditions where the individual ships are not trying to dodge incoming fire and can maintain stable relative positions. When in actual combat however, the typical ranges are short enough to make a sensor net unnecessary.

So how does it work in game terms? If the ships have identical PESA sensors, then the rule of thumb is that each doubling of the number of ships in the sensor net adds a plus 2 to the base sensor rating (indicating a larger aggregate tonnage of sensing equipment). Then you read the scan bonus from Table 2 and add it in to find the final scan rating for the sensor net. The table below can be used to determine the aggregate base sensor rating. If the ships do not mount identical equipment, then the situation is much more complicated and involves adding the mass of the individual sensors together and searching down the appropriate mass column on Table 1 to find the closest match to the total and determine their base aggregate rating before adding in the Table 2 bonus.

Table 3: Base Sensor Modifier
Number of Ships Modifier to Base Sensor Rating
2 +2
3 +3
4-5 +4
6-7 +5
8-11 +6
12-15 +7
16-23 +8
24-31 +9
32-47 +10


Two wings of Imperial Rampart-Class fighters (16 ships) are launched from their carrier to perform a sensor sweep of a system. The fighters form a loose formation (a few hundred miles of separation between ships) and link together to form a sensor net. All the fighters run a copy of Data Link to relay their raw sensor data via tight beam laser communicators and the fighter designated as the central processor also runs the Complexity-6 version of the Signal Processing software on its computer. The Ramparts each have a TL12 cockpit bridge that includes a scan-37 rated PESA sensor. (Nominal space range: 20,000 miles) Using Table 3, we find that combining the 16 passive sensors together yields an aggregate base sensor rating of 45 (37 + 8). Then, from Table 2, we determine that the modifier for a sensor net with 10-99 sub units is plus 2. So the final rating for the sensor net made by the 16 fighters is 47 (37 + 8 + 2), and the nominal space range (from Table 1) is 1,000,000 miles. When you factor in the size modifiers for large targets, the sensor net formed by the fighters can now detect other ships at ranges in the tens of millions of miles. So, if you use these sensor network rules, groups of fighters will be able to perform an important role in reconnaissance as the eyes of the fleet. Squadrons of capital ships, each mounting multi-ton PESA sensors, could also benefit by forming sensor nets. IMTU it is considered standard operating procedure to do so.


The two assumptions that all of the above sensor net rules are based on:

  1. The raw output from multiple small sensors can to combined and processed to at least match (and possibly exceed) the sensitivity of a single sensor of a size similar to the total of all the smaller sensors.
  2. Sensor input from multiple widely separated sources is superior to input from a single point source, for a variety of reasons detailed in the introduction above, and should act as a positive modifier to the scan rating of a sensor net.

I believe that there are a wealth of real world examples that support both of these assumptions and the rule system for sensor nets as presented above is a reasonable extension to the GURPS sensor rules.

The Last Word

The major objection I have encountered to the use of passive sensor nets as presented above is that Long Baseline Interferometry, while increasing the resolution of the processed image, will not improve the sensitivity of the component sensors and therefore should not increase the rating (range) of the sensor net above the rating of the components. However, if we project the capabilities our current passive detection equipment (Thermographic imagers and the like) into the future (say two GURPS tech levels or so), sensors of the size installed in the basic bridge of a starship should reasonably have sufficient sensitivity to detect the emissions of other starships at distances of hundreds of millions of miles. Since the effective ranges are multiple orders of magnitude smaller than that, I would conclude that while the passive sensors are sensitive enough to pick up the emissions, they lack the resolving power to distinguish ships or other targets at extreme ranges from the background noise. Therefore, increasing the resolution of the sensor by incorporating it into a sensor net should also improve its effective range.

As an alternative to sensor nets, the effective space range of passive sensors could be made much more realistic simply by using a ×1,000 rather than a ×10 multiplier for detection range in a vacuum. The main argument that I have heard for maintaining the un-realistically short detection ranges in GURPS Traveller is that they allow distance to cloak covert and/or illegal activity (including perhaps p****y) from detection. I personally find this line of argument less than compelling.

If you find any of the ideas or information presented in this essay to be useful, feel free to use it in your own campaign and reproduce it for personal use.


This page originally found at
Last updated: 25 July 2007