Uncertainty Theories and Multisensor Data Fusion

<p>INTRODUCTION ix</p>
<p>CHAPTER 1. MULTISENSOR DATA FUSION 1</p>
<p>1.1. Issues at stake 1</p>
<p>1.2. Problems 4</p>
<p>1.2.1. Interpretation and modeling of data 8</p>
<p>1.2.2. Reliability handling 10</p>
<p>1.2.3. Knowledge propagation 11</p>
<p>1.2.4. Matching of ambiguous data 12</p>
<p>1.2.5. Combination of sources14</p>
<p>1.2.6. Decision–making 16</p>
<p>1.3. Solutions 21</p>
<p>1.3.1. Panorama of useful theories 21</p>
<p>1.3.2. Process architectures 24</p>
<p>1.4. Position of multisensor data fusion 27</p>
<p>1.4.1. Peculiarities of the problem 27</p>
<p>1.4.2. Applications of multisensor data fusion 28</p>
<p>CHAPTER 2. REFERENCE FORMALISMS 31</p>
<p>2.1. Probabilities 31</p>
<p>2.2. Fuzzy sets 35</p>
<p>2.3. Possibility theory 39</p>
<p>2.4. Belief functions theory 43</p>
<p>2.4.1. Basic functions 44</p>
<p>2.4.2. A few particularly useful cases 47</p>
<p>2.4.3. Conditioning/deconditioning 49</p>
<p>2.4.4. Refinement/coarsening 50</p>
<p>CHAPTER 3. SET MANAGEMENT AND INFORMATION PROPAGATION 53</p>
<p>3.1. Fuzzy sets: propagation of imprecision 53</p>
<p>3.2. Probabilities and possibilities: the same approach to uncertainty 56</p>
<p>3.3. Belief functions: an overarching vision in terms of propagation 57</p>
<p>3.3.1. A generic operator: extension 58</p>
<p>3.3.2. Elaboration of a mass function with minimum specificity 61</p>
<p>3.3.3. Direct exploitation of the operator of extension 64</p>
<p>3.4. Example of application: updating of knowledge over time 66</p>
<p>CHAPTER 4. MANAGING THE RELIABILITY OF INFORMATION 71</p>
<p>4.1. Possibilistic view 72</p>
<p>4.2. Discounting of belief functions 73</p>
<p>4.3. Integrated processing of reliability 75</p>
<p>4.4. Management of domains of validity of the sources 77</p>
<p>4.5. Application to fusion of pixels from multispectral images 82</p>
<p>4.6. Formulation for problems of estimation 87</p>
<p>CHAPTER 5. COMBINATION OF SOURCES 91</p>
<p>5.1. Probabilities: a turnkey solution, Bayesian inference 92</p>
<p>5.2. Fuzzy sets: a grasp of axiomatics 94</p>
<p>5.3. Possibility theory: a simple approach to the basic principles 102</p>
<p>5.4. Theory of belief functions: conventional approaches 106</p>
<p>5.5. General approach to combination: any sets and logics 113</p>
<p>5.6. Conflict management 118</p>
<p>5.7. Back to Zadeh s paradox 122</p>
<p>CHAPTER 6. DATA MODELING 127</p>
<p>6.1. Characterization of signals 127</p>
<p>6.2. Probabilities: immediate taking into account 130</p>
<p>6.3. Belief functions: an open–ended and overarching framework 131</p>
<p>6.3.1. Integration of data into the fusion process 132</p>
<p>6.3.2. Generic problem: modeling of Cij values 135</p>
<p>6.3.3. Modeling measurements with stochastic learning 139</p>
<p>6.3.4. Modeling measurements with fuzzy learning 144</p>
<p>6.3.5. Overview of models for belief functions 148</p>
<p>6.4. Possibilities: a similar approach 153</p>
<p>6.5. Application to a didactic example of classification 157</p>
<p>CHAPTER 7. CLASSIFICATION: DECISION–MAKING AND EXPLOITATION OF THE DIVERSITY OF INFORMATION SOURCES 165</p>
<p>7.1. Decision–making: choice of the most likely hypothesis 166</p>
<p>7.2. Decision–making: determination of the most likely set of hypotheses 168</p>
<p>7.3. Behavior of the decision operator: some practical examples 171</p>
<p>7.4. Exploitation of the diversity of information sources: integration of binary comparisons 175</p>
<p>7.5. Exploitation of the diversity of information sources: classification on the basis of distinct but overlapping sets 179</p>
<p>7.6. Exploitation of the diversity of the attributes: example of application to the fusion of airborne image data 189</p>
<p>CHAPTER 8. SPATIAL DIMENSION: DATA ASSOCIATION 193</p>
<p>8.1. Data association: a multiform problem, which is unavoidable in multisensor data fusion 194</p>
<p>8.2. Construction of a general method for data association 197</p>
<p>8.3. Simple example of the implementation of the method 203</p>
<p>CHAPTER 9. TEMPORAL DIMENSION: TRACKING 211</p>
<p>9.1. Tracking: exploitation of the benefits of multisensory data fusion 211</p>
<p>9.2. Expression of the Bayesian filter 218</p>
<p>9.2.1. Statistical gating 218</p>
<p>9.2.2. Updating 219</p>
<p>9.2.3. Prediction 220</p>
<p>9.3. Signal discrimination process 221</p>
<p>9.3.1. Fusion at the level of each resolution cell 222</p>
<p>9.3.2. Fusion at the level of the validation gate 224</p>
<p>9.3.3. Overview of a practical implementation of the discrimination method 226</p>
<p>9.4. Extensions of the basic MSF 228</p>
<p>9.4.1. Data association 228</p>
<p>9.4.2. Joint tracking of multiple targets 229</p>
<p>9.4.3. Multi–model filtering 231</p>
<p>9.5. Examples of application 232</p>
<p>9.5.1. Extraction power 233</p>
<p>9.5.2. Handling of unfamiliar signatures 235</p>
<p>9.5.3. Tracking on spatially ambiguous observations 238</p>
<p>CONCLUSION 241</p>
<p>BIBLIOGRAPHY 249</p>
<p>INDEX 257</p>