| 摘要 |
A method for performing a statistical test on a device or radio channel, which has N outcomes in the form of N different events, whereby N is higher than two, with the following steps: measuring ns samples of the output of the device or radio channel under test, whereby occurs a specific number (na,nb,ne) of each event, defining a specific limit (L) for the test in a space spanned out by specific numbers of each event, erecting a N−1 dimensional likelihood distribution on several points of the limit (L), and constructing a threshold for fail of the radio channel or device and a threshold for pass of the radio channel or device by summing or integrating the N−1 dimensional likelihood distribution along unbroken paths parallel to the limit L until a predefined confidence level is reached.
|
