Fully Autonomous Toy Truck Managed by Eight Artificial Neurons
In a groundbreaking development, researchers have successfully controlled an RC truck using a Spiking Neural Network (SNN), a hardware-based artificial brain that mimics the timing and firing of biological neurons. The experimental vehicle, named the GSN SNN 4-8-24-2 Autonomous Vehicle, is a testament to the potential of this technology in the field of robotics.
Equipped with four proximity sensors at the front, front left, front right, and rear, the RC truck sends spike-based input signals to the external SNN. This network, composed of eight artificial neurons with 24 synapses, processes these inputs and generates control commands for steering and speed. These commands are then wirelessly sent back to the truck, enabling autonomous movement.
The SNN, implemented on breadboards with LEDs for visualising neural activity, processes discrete spike events rather than continuous signals. The strength and timing of these spikes influence the output decisions, making the system event-driven and low-level.
Interestingly, the same research group has also experimented with shaping the behaviour of a digital squid using a neural network. This suggests that SNNs or neuromorphic systems are not limited to vehicle control but can also be used for simulating and controlling bio-inspired agents in digital environments. However, detailed descriptions or results of the digital squid experiment are not widely available.
The GSN SNN 4-8-24-2 system provides a concrete example of this technology in action. A video showcasing the autonomous RC truck is available on the Global Science Network channel, with other videos delving deeper into the topics. The LEDs in the circuitry provide some insight into the system's functioning.
This innovation marks a significant step towards creating artificial systems that can mimic the complexity and adaptability of biological neural networks, opening up possibilities in robotics, artificial intelligence, and bio-simulations.
- The SNN technology, demonstrated in controlling the GSN SNN 4-8-24-2 Autonomous Vehicle, might find applications beyond vehicle control in the realm of science, particularly in the fields of space-and-astronomy, robotics, and technology, as it possesses the potential to simulate and control artificial agents.
- The integration of sensors, artificial intelligence, and neuromorphic systems in the GSN SNN 4-8-24-2 system signifies the fusion of various scientific disciplines, such as space-and-astronomy, technology, and artificial-intelligence, to create an autonomous vehicle that processes discrete spike events in a fashion akin to biological neural networks.
