Making Dronesophila Part 1

     Emulating the Drosophila nervous system was step one which I did successfully on a single laptop using functional groups. This works VERY well. 

The functional groups I created were:

Neuron Group    Synapse Count

central                     1455116

optic_lobes             1217922

visual_projection     268547

AN                             115292

Kenyon_Cell             92875

ALPN                     88302

CX                             85831

medulla_intrinsic     81621

visual_centrifugal     76532

DN                             58402

L1-5                     53746

ALLN                     34872

MBIN                     26501

bilateral                     25724

mechanosensory     19246

MBON                     18529

olfactory                     17689

LHCENT             12565

LHLN                     11217

gustatory                     7950

visual                     6400

unknown_sensory       4098

DAN                     3631

ALIN                     2301

ALON                     1992

ocellar                     1696

mAL                     1694

hygrosensory             1575

TuBu                     921

thermosensory     862

motor                     602

CSD                     164

optic                     108

pars lateralis             70

pars intercerebralis     19

sensory                     3

If you replace the Neuron Grouping labels into the boxes in the diagram above, each time a neuron fires, it knows whether it connects to a neuron within its functional group or a neuron outside of its functional group. If inside, it's just matrix functions, if outside, I use UDP to message to that group the neuron and weight that is being sent. 

The largest is the central complex. The list above is sorted by the number of synaptic connections with the largest number to the smallest. 

The above is the Central Nervous system of Drosophila. I am still waiting/investigating the Ventral Nerve Cord (VNC) neurons and will place them into their own functional groups when the data is completed. The VNC is congruent to a mammalian spinal column and peripheral nervous system, and is critical to the full extension of the Drosophila nervous system to the quadcopter.

With the entire nervous system, there are a number of sensing elements that need to be configured. Obvious sensory modalities are vision, pressure, thermal, smell (taste) but there are other modalities not as obvious. I want to use something like the Blue Robotics Navigator Flight Controller to give the emulation accelerometers and gyroscopes for orientation and magnetometers for head direction. The issue becomes what to connect these elements to within the Drosophila central nervous systems. I'm trying to figure this out?

Dronesophila Eyes

Once all these sensory modalities are connected, Dronesophila becomes self-aware! I do not say this lightly but its critical for flight that the robot understand and know its environment in order to allow the nervous system guide and navigate the quadcopter. Vision is very important but alone is not enough. This is one of the most common problems with current robotics in that there is too much emphasis on vision and ignores all other sensory modalities. 

The Central Complex is congruent to the mammal hippocampus. Place cells, grid cells, head direction cells and border cells are so important to our means to navigate domain agnostic environments. I am attempting to develop this same construct within Dronesophila's sensory modalities and emulated nervous system.