Feedforward layered perceptron neural networks seek to capture a system mapping inferred by training data. A properly trained neural network is not only capable of mimicking the process responsible for generating the training data, but the inverse process as well. Neural network inversion procedures seek to find one or more input values that produce a desired output response for a fixed set of synaptic weights. There are many methods for performing neural network inversion. Multi-element evolutionary inversion procedures are capable of finding numerous inversion points simultaneously. Constrained neural network inversion requires that the inversion solution belong to one or more specified constraint sets. In many cases, iterating between the neural network inversion solution and the constraint set can successfully solve constrained inversion problems. This paper surveys existing methodologies for neural network inversion, which is illustrated by its use as a tool in query-based learning, sonar performance analysis, power system security assessment, control, and generation of codebook vectors.

The Space Technology 7 Disturbance Reduction System (ST7-DRS) is a NASA technology demonstration payload that operated from January 2016 through July of 2017 on the European Space Agency's LISA Pathfinder spacecraft. The joint goal of the NASA and ESA missions was to validate key technologies for a future space-based gravitational wave observatory targeting the source-rich milliHertz band. The two primary components of ST7-DRS are a micropropulsion system based on colloidal micro-Newton thrusters (CMNTs) and a control system that simultaneously controls the attitude and position of the spacecraft and the two free-flying test masses (TMs). This paper presents our main experimental results and summarizes the overall the performance of the CMNTs and control laws. We find that the CMNT performance to be consistent with pre-flight predictions, with a measured system thrust noise on the order of 100 nN/ √ Hz in the 1 mHz ≤ f ≤ 30 mHz band. The control system maintained the TM-spacecraft separation with an RMS error of less than 2 nm and a noise spectral density of less than 3 nm/ √ Hz in the same band. Thruster calibration measurements yield thrust values consistent with the performance model and ground-based thrust-stand measurements, to within a few percent. We also report a differential acceleration noise between the two test masses with a spectral density of roughly 3 fm/s 2 / √ Hz in the 1 mHz ≤ f ≤ 30 mHz band, slightly less than twice as large as the best performance reported with the baseline LISA Pathfinder configuration and below the current requirements for the Laser Interferometer Space Antenna (LISA) mission.

No abstract

Near-optimal feedback controls for minimax-range pursuit-evasion problems between two constant-thrust spacecraft are generated by periodically resolving the differntial game based on the actual system state using a modified version of a first-order differential dynamic programming algorithm. Compared to a previously developed technique that requires the backward integration of a matrix Riccati differential equation, this new technique can be implemented in real time much more easily, and it requires only a rough estimate of the optimal controls to start it instead of a complete two-point boundary-value problem solution. Numerical results are presented which illustrate the advantages and limitations of this new technique.

No abstract