Abstract:
Robotic swarms can benefit from means of communications between agents to perform
most collaborative tasks. A communication method that scales linearly with the swarm
size is required as the number of agents increases significantly. We propose a method that
communicates inherently with proximity, as opposed to traditional fully connected communications.
Also knowing swarm state information can improve the efficiency of the swarm.
Visible Light Communication (VLC) using a camera and LEDs solves two problems for
swarms at once, communication and the location of the sending agent. Information from
the frequency of a blinking LED can be inferred over time. Parsing data from many agents
in a single agent's frame simultaneously is possible due to using a camera and segmenting
the frame. The location of the sending agent can be extracted in the local frame of
the receiving agent with the camera frame and the centroid of the light source. In this
method of communication, the receiver's communication rate is scaled with the number of
agents in \frame". When the camera is not being used to communicate, it can be used
to collect information from the environment or usage for other tasks. Methods to perform
communication and localization using LEDs paired with Digital and event-based cameras
are presented. Methods to improve the data rate using Frequency Shift Keying(FSK) and
Multiple Frequency Shift Keying(MFSK) were presented and tested for their effectiveness.
Results showed a low bit-error rate using a camera while inferring spatial locality of the
sending agent, showing the feasibility for VLC for robotic swarms.
