Autonomous Toy Truck Regulated by Eight Synthetic Neurons
In a groundbreaking demonstration, the Global Science Network has showcased an RC truck named the GSN SNN 4-8-24-2 Autonomous Vehicle. This experimental vehicle is controlled by a Spiking Neural Network (SNN), marking a significant step forward in the field of artificial intelligence.
The SNN hardware, housed in breadboards, comprises eight artificial neurons and 24 synapses. These neurons process discrete spike events from four proximity sensors (front, front-left, front-right, and rear) surrounding the truck. The timing of these spikes is crucial, as it encodes when and how strongly neurons activate. This allows the network to interpret sensor data, such as detecting obstacles, and make decisions on whether to move forward, backward, or turn.
The output commands from the artificial brain are wirelessly sent to the RC truck's controller, which has been modified to receive electronic signals that simulate pressing buttons for movement. This setup enables the truck to navigate autonomously without crashing. LEDs on the vehicle and the breadboards provide visual feedback of spiking activity in real time, offering an insight into the neural processing.
In addition to the RC truck project, the Global Science Network has also produced videos exploring the possibility of shaping the behavior of a digital squid using a neural network. These videos, if you're interested, can be found on their channel or platform.
For a comprehensive understanding of the RC truck controlled by an SNN, you can watch a detailed demonstration and explanation video on the Global Science Network's YouTube channel. The video provides an in-depth look at the technical setup and live autonomous driving behaviour of the vehicle [3].
| Aspect | Details | |----------------------------|-------------------------------------------------------------------| | SNN Hardware | 8 neurons, 24 synapses, breadboard circuit | | Sensors | 4 IR proximity sensors (front, front-left, front-right, rear) | | Input Processing | Sensor readings trigger neuron spikes based on thresholds | | Output Commands | Wireless signals control the RC truck’s motors (forward/back/turns) | | Behavior | Autonomous obstacle avoidance and navigation | | Visualization | LEDs reflect neuron firing activity | | Digital Squid Videos | Neural network-shaped behavior videos exist by the same creators |
[1] Global Science Network's Digital Squid Videos [2] LED Visual Feedback in the GSN SNN 4-8-24-2 Autonomous Vehicle [3] Detailed Demonstration and Explanation Video of the GSN SNN 4-8-24-2 Autonomous Vehicle on the Global Science Network's YouTube Channel
The Global Science Network's RC truck is controlled by an SNN hardware, which includes eight neurons and 24 synapses, processing sensor data from four proximity sensors. This data is encoded through spike events, allowing the network to interpret the surroundings, make decisions, and navigate autonomously.
The SNN's potential applications extend beyond the SNN 4-8-24-2 Autonomous Vehicle, as the same network was also used to manipulate the behavior of a digital squid in videos produced by the Global Science Network.
