Bioinspired Navigation Based on Distributed Sensing in the Leech
277
Various computational models have been implemented to simulate the neural
response of neurons in animals [11–17]. The model proposed in [11] generates
elliptical representations of spike rate distributions for two different populations
of neurons: neurons that are sensitive to magnetic stimuli, and neurons that
are sensitive to visual stimuli. The two representations are integrated to produce a resultant effective ellipse which computes the navigational behavior of
the simulated agent. The results of this model matched the results of real-world
animal behavior experiments where birds were given different visual and magnetic stimuli in a laboratory. The approach was later adapted by [13] to study
magnetoreception and navigation.
In this study, the elliptical approach of [11] (referred to as the elliptical model)
was adapted to model the leech’s nervous system response to both mechanical
and visual stimuli. Once the elliptical model was able to generate biologically
relevant neural activity, a behavioral algorithm was used to transform the neural
responses into behavioral actions. A computer simulated agent then used these
actions to navigate towards the center of a given stimulus. We analyzed the trajectories of the agent to verify that the model exhibits similar behavior to that of
a leech. From a biological standpoint, our model can serve as a complementary
testing tool alongside real-world leech experiments to help discover the underlying mechanisms and principles of distributed sensing, sensor placement, and
multimodal sensing. From an engineering standpoint, our model can be used in
the development of man-made sensing, processing, and information systems and
tools that are more precise, autonomous, and efficient.
2 Materials and Methods
This work uses an agent-based simulation in conjunction with the elliptical model
to conduct all trials and experiments. Figure 2 illustrates a high-level overview
of how our model and agent-based simulation works. Each section explains a
piece of our system. Section 2.1 outlines how mechanical and visual stimuli are
generated. Section 2.2 how spike rates are generated based on different stimuli.
Sections 2.3 and 2.4 how the elliptical model is used to represent the leech’s
nervous system response. Sections 2.5 and 2.6 describe the navigational environment, and the parameters used in the experiments.
2.1 Stimuli Generation
In this study, the mechanical stimulus is simplified to be represented by a frequency and intensity value. Since the mechanical intensity of physical water
waves decreases as they move across space, the intensities are modeled as a function of the displacement between the agent’s location and the stimulus origin.
277
Various computational models have been implemented to simulate the neural
response of neurons in animals [11–17]. The model proposed in [11] generates
elliptical representations of spike rate distributions for two different populations
of neurons: neurons that are sensitive to magnetic stimuli, and neurons that
are sensitive to visual stimuli. The two representations are integrated to produce a resultant effective ellipse which computes the navigational behavior of
the simulated agent. The results of this model matched the results of real-world
animal behavior experiments where birds were given different visual and magnetic stimuli in a laboratory. The approach was later adapted by [13] to study
magnetoreception and navigation.
In this study, the elliptical approach of [11] (referred to as the elliptical model)
was adapted to model the leech’s nervous system response to both mechanical
and visual stimuli. Once the elliptical model was able to generate biologically
relevant neural activity, a behavioral algorithm was used to transform the neural
responses into behavioral actions. A computer simulated agent then used these
actions to navigate towards the center of a given stimulus. We analyzed the trajectories of the agent to verify that the model exhibits similar behavior to that of
a leech. From a biological standpoint, our model can serve as a complementary
testing tool alongside real-world leech experiments to help discover the underlying mechanisms and principles of distributed sensing, sensor placement, and
multimodal sensing. From an engineering standpoint, our model can be used in
the development of man-made sensing, processing, and information systems and
tools that are more precise, autonomous, and efficient.
2 Materials and Methods
This work uses an agent-based simulation in conjunction with the elliptical model
to conduct all trials and experiments. Figure 2 illustrates a high-level overview
of how our model and agent-based simulation works. Each section explains a
piece of our system. Section 2.1 outlines how mechanical and visual stimuli are
generated. Section 2.2 how spike rates are generated based on different stimuli.
Sections 2.3 and 2.4 how the elliptical model is used to represent the leech’s
nervous system response. Sections 2.5 and 2.6 describe the navigational environment, and the parameters used in the experiments.
2.1 Stimuli Generation
In this study, the mechanical stimulus is simplified to be represented by a frequency and intensity value. Since the mechanical intensity of physical water
waves decreases as they move across space, the intensities are modeled as a function of the displacement between the agent’s location and the stimulus origin.