Conclusion
Multitrack Fusion

Clearly, there are many other track merging algorithms based on other, more sophisticated clustering (i.e. unsupervised classification) algorithms, making it possible, for example, to make the distance between representatives robust to the laws followed by the errors and to develop other stopping tests.

Nevertheless, the approach presented here has the advantage of ensuring a certain "intellectual" uniqueness between two fields that are often split up, namely upstream processing (known as TS) and downstream processing (known as TI). This helps teams to understand each other, a key factor in the successful implementation of such monitoring systems. What's more, this approach does not require additional training in less standard disciplines such as expert systems, neural networks, new theories of "uncertainty" and many other approaches in the field of artificial intelligence...