**3.4 Basic concepts of evidence theory**

D-S evidence theory [11, 12] adopts mathematical reasoning to perform fusion calculations of inexact and incomplete information. In the D-S evidence theory fusion algorithm, the recognition framework is the framework of the whole judgment; the Basic Probability Allocation is the basis of fusion; the combinational rule is the fusion process, and the trust function and likelihood function are used to express the upper and lower limits of support strength interval of fusion conclusion to a hypothesis.

1.Recognition framework

Θ is a mutually exclusive non-empty finite set, which is known as recognition framework. It consists of N nonintersecting sets of *w*1, *w*2, *w*3*:* … *wN*, and there are N possible hypotheses in this recognition framework. The task of the evidence theory fusion algorithm is to estimate the trust level to each possible hypothesis.
