No abstract

We describe a new way of collecting data from Bomem DA3 Fourier transform spectrometers, bypassing the original computers used by these instruments, which are obsolete and more prone to failure than other parts of the system. We collect the interferogram, helium-neon reference laser, and a trigger marking zero path di↵erence with a modern computer as a function of time at a high data rate, as opposed to as a function of position along the scanning arm of the interferometer. The interferogram is then reconstructed as

Bands of the g 6Φ-X 4Δ, g 6Φ-A 4Π, g 6Φ-a 6Δ, and g 6Φ-b 6Π electronic transitions of iron monodeuteride (FeD) have been measured in laser excitation and in dispersed fluorescence. The molecules were produced both in a cold supersonic molecular jet source and in a chemical reaction between iron pentacarbonyl [Fe(CO5)] and a microwave discharge of argon and hydrogen gases. Dispersed fluorescence from the latter source was detected at high resolution with a Fourier transform spectrometer, yielding a large number of the transitions observed. The data reveal that FeD experiences strong interstate couplings that compromise fitting of the data with traditional Hamiltonians but that the problem is less severe than in corresponding spectra of FeH. This work greatly expands the available data on FeD, which were previously characterized only through the F 4Δ-X 4Δ spectrum and pure rotational data in the ground state.

This paper presents a deep-learning based CPP algorithm, called Coverage Path Planning Network (CPPNet). CPPNet is built using a convolutional neural network (CNN) whose input is a graph-based representation of the occupancy grid map while its output is an edge probability heat graph, where the value of each edge is the probability of belonging to the optimal TSP tour. Finally, a greedy search is used to select the final optimized tour. CPPNet is trained and comparatively evaluated against the TSP tour. It is shown that CPPNet provides near-optimal solutions while requiring significantly less computational time, thus enabling real-time coverage path planning in partially unknown and dynamic environments.