New Experimental Capabilities¶

Simultaneous Full Field-of-View Imaging¶

Simultaneous Full Field-of-View (FOV) Imaging is now possible with Borealis! The current implementation uses phase modulation across the transmitting antennas of the main array. Two configurations are supported right now: transmission with all 16 main antennas, and transmission with 8 adjacent antennas of the main array. These two modes are implemented in the experiments full_fov.py and full_fov_2freq.py, respectively, with the latter using 8 antennas for each of two frequencies which transmit simultaneously.

To create your own Full FOV experiment using 8- or 16-antenna phase modulation, a utility function is present in superdarn_common_fields.py. To use this function, set

slice_dict['tx_antenna_pattern'] = scf.easy_widebeam

in your experiment slice dictionary.

More generally, you can define your own power- and phase-modulation across the transmitting antennas. Define a function with the following signature

def fn_name(frequency_khz, tx_antennas, antenna_locations):
    """Defines the power and phase modulation for each transmitting antenna.

    Parameters
    ----------
    frequency_khz: int
        Frequency in kHz.
    tx_antennas: list
        List of transmitting antennas for the slice.
    antenna_locations: np.ndarray
        Location of each antenna, in [x, y, z] coordinates. Shape [num_antennas, 3]

    Returns
    -------
    np.ndarray of shape (num_beams, num_main_antennas). Each element should be a complex number
    with magnitude <= 1.0 defining the power and phase for that antenna.
    """

Then, assign the function to tx_antenna_pattern in your slice dictionary. It is important that the first dimension of the returned array matches the first dimension of rx_beam_order in your slice dictionary, and that num_main_antennas matches the number of main antennas in the config file.

Custom beamforming of the results measured during a full field of view experiment is possible through defining the rx_antenna_pattern field in the full field of view experiment. A custom function can be written in the experiment and passed to borealis

beamforming_function(beam_angle, freq, antenna_locations):

...

slice_dict['tx_antenna_pattern'] = beamforming_function

The function should expect to receive beam angles, operating frequencies, and antenna locations, This function will be called for both the rx signals from the main array and the interferometer array. The return is expected to be the desired phase for beamforming each antenna, and should be of size [beam_angle, antenna_count]. The magnitude of each entry should be less than or equal to 1.

Bistatic Experiments¶

Bistatic experiments are now supported in Borealis. A multi-purpose bistatic experiment is defined in bistatic_test.py, and can be used at both transmitting and receiving radar sites. This experiment uses command-line arguments to control its behaviour, making it flexible and configurable.

By default, the bistatic_test experiment will transmit a full FOV pattern like full_fov.py on a single frequency. To operate a radar in a bistatic listening mode, the argument listen_to=[three-letter radar code] must be passed to radar_control via steamed_hams.py. This will look something like this

radar@borealis~$ ./steamed_hams.py bistatic_test release discretionary --kwargs_string "listen_to=rkn"

This invocation will trigger the radar to receive only, tuning to common-mode frequency 1 for RKN defined in superdarn_common_fields.py. Additionally, it will trigger an imaging mode, receiving signals on all 16 beams simultaneously. If the listening radar specified is the same as the radar that you are running the experiment on, the experiment will default to a listening-only mode on the radar’s common-mode frequency 1.

For further control over the transmitting characteristics, an additional keyword argument beam_order is supported. This controls the tx_beam_order field of the slice dictionary, and allows for traditional beams to be used. The beam_order value must be formatted as a list of numbers, such as 0,1,2,3-5,2,9, with ranges being parsed to include all numbers in between and both endpoints. Therefore, for this example, beams used would be 0, 1, 2, 3, 4, 5, 2 and finally 9. Repeated beams are valid.

The arguments beam_order and listen_to are mutually exclusive.

You can define your own bistatic experiment, with very few restrictions. It is highly recommended that the field align_sequences is set to True in your experiment slice dictionary, which will send out the first pulse in each sequence within 1us of each 0.1 second boundary. Without this field set in the experiments of both radars in the bistatic link, there is no guarantee of timing synchronicity and the data will likely be useless. Additionally, it is recommended that the experiments running at both the transmit and receive radars are both using the same scanbound. This will make it much easier to compare data from the transmit and recieve sites as the averaging periods should line up exactly. Lastly, it is recommended that you check the data files for both radars afterwards and ensure that the gps_locked flag is True for all times. If not, the clock may have drifted, and the sqn_timestamps field may be inaccurate.

Clear Frequency Search (Experimental)¶

An experimental implementation of clear frequency searching is now supported in borealis to determine a transmit frequency within a specified band. In an experiment slice, the freq parameter should be unset, and the cfs_range parameter set to a two element list containing the lower and upper frequency limits of the CFS band:

slice['cfs_range'] = [11000, 11300]  # Lower and upper freq limit in kHz

Multiple experiment slices within an averaging period can be configured to set a transmit frequency from the CFS as long as each slice has the cfs_range set. Each slice can choose any band within the transmit and receive bandwidth of the system. Be aware when choosing a cfs_range that if the range has any part within +/- 50kHz around the txctrfreq or rxctrfreq a warning will be raised as no tx frequency can be chosen that is within 50kHz of the center frequencies. The user should be aware of any restricted bands within the desired range, as CFS will exclude restricted bands when selecting transmit frequencies.

Additionally, if a cfs_range with a band greater than 300kHz is desired, the user will need to design a custom decimation scheme for the CFS analysis, as the default is designed only for bands of 300kHz or smaller. All CFS slices with CONCURRENT or SEQUENCE interfacing must have the same decimation scheme, power threshold, stable time, duration, and frequency resolution (parameters defined below).

The following parameters can be set for a CFS slice:

Clear Frequency Search Experiment Parameters¶
CFS Parameter	Default Value	Description
`cfs_range`	`None`	Sets the band to search. Must be set to perform CFS
`cfs_duration`	90 ms	Determines how long the CFS sequence will listen for
`cfs_scheme`	`create_default_cfs_scheme()`	Decimation scheme used in analyzing the CFS data. The default scheme is designed for bands of 300kHz or less
`cfs_stable_time`	0 s	Sets a minimum amount of time during which CFS will not change the frequency of a CFS slice after it has been set. By default the CFS measurements will only be made once the stable time has elapsed.
`cfs_always_run`	False	When true, CFS measurements will be taken every averaging period even if the `cfs_stable_time` has not elapsed. Measurements taken before the stable time has elapsed will not be used to set new transmit frequencies but the measurements will still be written to file.
`cfs_pwr_threshold`	`None` (dB)	Sets a threshold power difference that a CFS scan must exceed before a frequency is switched. If another frequency is lower in power than the current frequency was when set by the threshold or if the current frequency power has increased by more than the threshold, then CFS will set a new frequency for the CFS slices. When set the threshold value must be greater than zero.
`cfs_fft_n`	512	Sets the number of samples used in the FFT during CFS processing. Determines the frequency resolution of the CFS, where the resolution is `res = (rx_rate / total decimation rate) / N`
`cfs_freq_res`	`None` (Hz)	Cannot be set if `cfs_fft_n` is set. Defines the desired frequency resolution of CFS in Hz. The `cfs_fft_n` parameter is then set to the nearest integer value of `N` calculated as `N = (rx_rate / total decimation rate) / res`

When a CFS slice is to be run during an averaging period, the first sequence of the averaging period is used to listen for the length of time specified by cfs_duration. The data from this measurement is beamformed based on the beam transmit direction and the center frequency of the cfs_range.The beamformed results are processed to evaluate the frequency spectrum of the collected data and select the least noisy frequencies for transmission in that beam direction. The analysis results are also recorded to any generated rawacf, antennas_iq, and/or bfiq files. The frequency of each beam is set and tracked by the CFS independently.