The concept of the transfer of consciousness of the worm in the robot

The main task is to realize the nervous system of the roundworm c. elegans in the body of the robot in such a way that all behavior is controlled by this neural network.

Why c. elegans?

A widely known and most well-studied organism with detailed information about neurons and a complete description of the structure of connections in the nervous system.


Figure 1. Nematode model in NeuroConstruct .
')

What is the fan?

The idea of ​​transferring the nervous system of this nematode to a robot is not new and has already appeared many times.
For example, popular connectom implementations in Raspberry Pi GoPiGo or Lego EV3 robots.

Usually, a similar problem is solved with the help of an adder with selected weights of inhibitory connections, which allow changing the trajectory when an obstacle is detected.
Which is definitely a victory.

But, for the sake of interest, (and the availability of free time), I wanted to understand the morphology and behavior of nematodes a little deeper. And the very idea of ​​implementing a simplified biological model for controlling a robot does not cease to fascinate.

Behavior

Based on real behavior c. elegans would be nice to simulate the following activity:
1. Moving forward.
2. Change of direction in case of detection of an obstacle.
3. Turns left / right.

By connect

A nematode connection is a simple list of neurons and the connections between them with already defined weights and the type of connection. In general, we are only interested in exciting and inhibiting types of relationships.
You can familiarize yourself with the links here or here or in the Celegans NeuroConstruct distribution kit .

Unfortunately, it’s not so easy to find the inhibitory neurons in the list. For this, it was necessary to study both the neurons themselves and the neurotransmitters. Here is an example of a braking neuron (inhibitory neuron).
(It is worth noting that I still have many questions, for example, why there are so few inhibitory neurons in the connection, although in the articles they write about the “inhibitory effect” of one group of neurons on others. For example, RIM neurons inhibit AVB, but the connection between them not inhibitory. And so on. There are also neurons whose role is not entirely clear or the nature of the connections is not clear.)

Neuron

The basis of the whole model should be an artificial neuron. For more drama, it was decided to use the spiking model of the Izhikevich neuron , which allows not only to visualize neural activity, but also has a number of interesting properties, such as synaptic currents, membrane potentials, and generally simulate neural phenomena that are not close to real ones.

To the realization of the neuron Izhikevich, Hebb's differential training was added to strengthen the connections between neurons.


Figure 2. Visualization of the neuron spike model. Membrane potentials.

Neural network and behavior

A neural network based on a connectom is now easily organized.
Based on the expected behavior, we find groups of neurons, the stimulation of which will lead to the required activity.
For example, the activation of sensory neurons responsible for the detection of food leads to movement:
ADFL, ADFR, ASGL, ASGR, ASIL, ASIR, ASJL, ASJR, AWCL, AWCR, AWAL, AWAR.

It is usually said that the movement is forward, but in reality the direction of movement may suddenly change.

When stimulating sensory neurons responsible for detecting an obstacle , the direction of movement changes :
ASHL, ASHR, FLPL, FLPR, OLQDL, OLQDR, OLQVL, OLQVR, IL1VL, IL1VR, IL1L, IL1R, IL1DL, IL1DR.

Indeed, when stimulating the search receptors for food, both the rhythm and the specific pattern of activation on muscular neurons are noticeable.

But the excitation of the nasal receptors, which should lead to a change in the direction of movement, also activates the muscular neurons with a characteristic pattern.

 d: - - 1 - - - - - - - 1 1 1 - - - 1 - - - - - - -  
 v: - - - - - - - - - - - - - - - - - - - - - - - - -  
 v: - - 1 - - - - - - - - - - - - - - - - - - - - - -  
 d: 1 - 1 - - - - - - - 1 1 1 - - - 1 - - - - - - -  

D and v means dorsal and ventral muscles, respectively. Muscle neuron activation pattern using the neuron spike model.


Figure 3. The joy of the eyes or the difference between the activity of the abdominal and dorsal muscular neurons. In the dynamics forms a wave-like movement.

Muscular neurons are also active when moving forward as they are moving backward.

But how to determine in which direction the movement is going or at least the moment of changing the direction of movement?

Here it is described that the forward movement is triggered by a signal from AVB and PVC neurons to B neurons, and backward movement from AVA, AVD and AVE to A neurons.

It is also reported that the forward and backward movements are different kinds of activity and are caused by different areas of the nervous system. Although it is noted that these areas interact with each other. And the neurons responsible for moving forward play some role when moving backwards. Those. active when driving.
But here it is shown that B neurons are active when moving forward, when A neurons are active when moving backwards.


Figure 4. Changes in the activity of VA and VB neurons and their dependencies. More details here .

Essentially means that the motor neurons of VB and VA should not be simultaneously active while moving.

This gave us only the fact that we need to watch the activity of VA neurons.

After analyzing the activity in stimulating food neurons (food sensors) and nose neurons (nose touch sensors), it turned out that in the latter case, the frequency of spikes increases.


Figure 5. The average activity of VA neurons. At around 900000 you can see how the frequency of the pulses has increased. This is the moment of the start of nasal neuron stimulation.

This allowed us to find a way when to change the direction of movement:

In general, this is the calculation of the average frequency of oscillation during movement and obstruction. If at the moment of time the average frequency is closer to the movement, then the movement occurs, if closer to the obstacle, the direction must be changed accordingly.


Figure 6. Stimulation of the nose stops at around 980,000, which leads to a sharp drop in the average activity of both VA and VB neurons.


Figure 7. We hope to change the direction of movement from the moment of 980000.

Simulation

For the simulation of the robot, the open source project Enki was chosen, which provides the ability to simulate robots on a flat surface with basic physics support.

Enki includes the implementation of several academic robots, of which the choice fell on the E-puck , to implement on its basis a connection.

E-puck has a range of infrared sensors.


Fig 8. Diagram of the sensors on the robot.

Sensors 0, 7, 3, 4 were selected to stimulate neuronal obstacles (nose touch).

As a result, the robot was able to change its direction while stimulating a given group of neurons and, as a result, all three patterns of behavior were achieved: moving forward, changing direction and turning.


Fig. 8. Simulation of an extended version of the E-puck robot with integration of a realized connection based on the spike model of the Izhikevich neuron.


Figure 9. Handsome and well done E-puck robot.

Conclusion

Thanks to those who have mastered it, but the truly brave and desperate are also familiar with the implementation of this realization , and the title humor is taken from here .