DETECT AND CONFIRM
E
very year, the U.S. government pours billions of dollars of taxpayer money into the development of sensors and sensing technologies. However, while the government spends lots of money developing "new
and improved" sensing methods, comparatively little attention is given to developing holistic methods to fuse data from exist- ing sensors, to achieve results whereby the whole is greater than the sum of its parts. While there are probably many reasons for this, it is at least partly because sensors are typically developed to monitor a single isolated detection thing called an "observable," rather than to work in a complementary way with other sensors to monitor a continuum of observables.
Additionally, people tend to think that sensor fusion is some- thing that needs complex mathematical formalism (i.e., it has to be really hard), so that it can only be successfully accomplished with lots of brain (and computing) power.
Tis is contrary to my experience, as I have found sensor fusion to be innate and practical, something that we humans perform many times, every day. With this in mind, my goal is to demys- tify sensor fusion, at least a little. To do this, consider what I call the "locate and confirm" sensor fusion method, which is one of the more practical ways that sensor data can be fused to yield improved detection results.
LOCATE AND CONFIRM Sensors detect. With simpler sensors, sometimes that's all they do. Tink of a thermometer. It detects temperature. More com- plex sensors also detect, but because of the more detailed and complete way that they detect, they can also confirm or rule out. Tis is why it is helpful to "fuse" information gathered from sensors.
Te locate-and-confirm method helps because it can allow for very low false-alarm rates and because it is fairly simple: It typically takes only a two-step detection process and some very basic reasoning methods. In short, one sensor detects, or locates, something that is not right—an anomaly—and the second sen- sor confirms a smart guess that can be made either by a person or a computer, or rules it out. Being able to quickly rule out some- thing bad and lower the false-alarm rate is just less expensive.
Imagine a mother with her young child. Te child appears to be sick: Is the cause a routine cold or flu, or is it a highly contagious childhood disease, such as mumps or measles? To find out, the mom takes her child to the family doctor.
What the doctor will search for are, in terms of sensors, anoma- lies. Normally the child feels well, but now does not. Detecting the anomalies will help determine why, and in turn will help with the diagnosis.
Te first step is a routine exam, or scan. For example, the nurse might take the child's temperature. Is there a fever? Te nurse also might scan the child's ears—are they red or swollen? Is the child's throat red, sore or swollen?
Even if the answer to each of these questions is yes, that is not a diagnosis, but the doctor can use these detected anomalies to come up with a smart guess about why the child is sick.
Like the doctor, a sensor user is trying to find anomalies. For example, when attempting to detect airborne biological agents, a sensor might look for an anomaly in the form of an abnormally large concentration of aerosolized particles, in the particle size range where they are optimally taken up by the lungs. Similarly, when attempting to detect buried land mines, you might employ a sensor that could detect anomalies in the form of buried objects that are moderately large and have a metallic signature.
Most of the time, anomaly detection is simple, straightforward and relatively inexpensive. Tere are frequently anomaly detec- tors that work well for lots of uses. However, while finding an anomaly may seem a cheap way of coming to a quick answer, the anomaly alone is not the answer. For anomaly detection to be useful in providing the whole picture, the detection process must set the stage for the user to determine the cause for the anomaly.
If the doctor suspects a throat or ear infection, he or she will try to confirm that guess—strep throat, maybe—with a throat swab and culture that then goes to the lab for confirmation—or rule it out. In the same way, the sensor user wants to be able to figure out what caused the anomaly.
A SIMPLE, POWERFUL TOOL Generally, detectors used in the confirmation step of the locate- and-confirm method are more difficult to find than anomaly detectors, and more expensive. Also, they can be harder to use, especially in the circumstances of a military operation. (Simi- larly, it's easy for the doctor, in his home office, to take the swab of the child's throat and send it to a lab for confirmation, but much more difficult if the doctor is working in a remote village without electricity or a nearby lab.) Still, even with cost and dif- ficulty, confirmation sensors may be worth it because they can lower false-alarm rates.
76
Army AL&T Magazine
April-June 2016
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92 |
Page 93 |
Page 94 |
Page 95 |
Page 96 |
Page 97 |
Page 98 |
Page 99 |
Page 100 |
Page 101 |
Page 102 |
Page 103 |
Page 104 |
Page 105 |
Page 106 |
Page 107 |
Page 108 |
Page 109 |
Page 110 |
Page 111 |
Page 112 |
Page 113 |
Page 114 |
Page 115 |
Page 116 |
Page 117 |
Page 118 |
Page 119 |
Page 120 |
Page 121 |
Page 122 |
Page 123 |
Page 124 |
Page 125 |
Page 126 |
Page 127 |
Page 128 |
Page 129 |
Page 130 |
Page 131 |
Page 132 |
Page 133 |
Page 134 |
Page 135 |
Page 136 |
Page 137 |
Page 138 |
Page 139 |
Page 140 |
Page 141 |
Page 142 |
Page 143 |
Page 144 |
Page 145 |
Page 146 |
Page 147 |
Page 148 |
Page 149 |
Page 150 |
Page 151 |
Page 152 |
Page 153 |
Page 154 |
Page 155 |
Page 156 |
Page 157 |
Page 158 |
Page 159 |
Page 160 |
Page 161 |
Page 162 |
Page 163 |
Page 164 |
Page 165 |
Page 166 |
Page 167 |
Page 168 |
Page 169 |
Page 170 |
Page 171 |
Page 172 |
Page 173 |
Page 174 |
Page 175 |
Page 176 |
Page 177 |
Page 178 |
Page 179 |
Page 180 |
Page 181 |
Page 182 |
Page 183 |
Page 184 |
Page 185 |
Page 186 |
Page 187 |
Page 188 |
Page 189 |
Page 190 |
Page 191 |
Page 192 |
Page 193 |
Page 194 |
Page 195 |
Page 196 |
Page 197 |
Page 198 |
Page 199 |
Page 200 |
Page 201 |
Page 202 |
Page 203
