wiki:2015/collisiondetection15

Since we are still working on getting input from the pushbot, we can tune the network with the data captured from a static DVS retina. Please download them from the attachments. The data is named like this: d0_p0_b_0: moving direction 0 degree, position in the middle, backward(/forward), sequence number. There are .aerdat raw files and .spikes files which can be used as spike source array in SpiNNaker.

You can use Matlab to read the .aerdat file and to play with it. To display we can use display_record. cin = dat2mat2('xxx.aedat', 128); % the resolution is 128*128 display_record( file_name, pause_time, dsize); % pause_time = 0.1s dsize=128

The spike source array works like this:

def read_spikefile(file_name, nNeurons):

spike_array = [[] for x in range(nNeurons)] with open(file_name) as f_spike:

for line in f_spike:

cut_index = line.find(';') time_stamp = int(line[0:cut_index]) neuron_list = line[cut_index+1:-1].split(',') for neuron in neuron_list:

neuronID = int(neuron) spike_array[neuronID].append(time_stamp)

return spike_array

spikes = read_spikefile('xxx.spikes', nNeurons) spikeArray = {'spike_times': spikes} populations.append(p.Population(nNeurons, p.SpikeSourceArray?, spikeArray,

label='inputSpikes'))

Last modified 4 years ago Last modified on 05/04/15 16:13:56

Attachments (5)