After many months of searching in the South Indian Ocean, public pressure has finally forced Inmarset to release the data that was used to construct the above and below the sea search for the missing plane. After more than a month of exhaustive searching , despite optimistic CNN reports nothing has been found. Today (5/27/2014) Inmarset's eagerly awaited report was released. This writer, along with other aerospace engineers, was anxious to see if Inmarset's claim of having "discovered a new tracking algorithm" was valid. Alas, in addition to wasting valuable search resources, Inmarset's claim is proving to be pure BS.
Back to basic fundamentals: In order to determine an A/C's position one needs 3 orthogonal measurements, i.e., XYZ or R and azimuth and elevation. R (range) and Doppler which is actually Range-Rate are insufficient. Range measurements from multiple satellites could also be used to determine the A/C's position, but all measurements were from a single satellite. It would appear that Inmarsat has lost sight of the fundamentals. Too bad none of CNN's paid "experts" had any basic knowledge of geometry. Whether intentional or not, Inmarset's marketeers were eager "to sell their product" and used gullible " CNN reporters to spread their BS.
As to Inmarset's claim to have discovered a new tracking algorithm; Any legitimate algorithm can be explained via equations or a flow chart. If provided, and they weren't, any sold engineering graduate would be able to understand their "algorithm." Inmarset marketeers have successfully fooled many non technical managers and reporters.
However, this still leaves the mystery of the acoustic ping data which appears to verify that they were searching in the correct place. The raw, time series data from the underwater Bluefin detector should also be made public. Outside Sonar signal processing experts should be brought in to analyze and explain the results. Oceanographers need to explain the shift in ping frequencies. In addition, from my limited experience in this area, I know that spectral (FFT) processing of limited time series data can yield false lines.