|
"the Fold: episode III (Reshaped agents: ά4, ά5, ά6)", robots, 3d printed elements, MISBkit electronics, video projection livestream, Expanded Arts Studio, ASU, Tempe, AZ 2026 |
Reshaped Agents are a continued investigation of the robotic agents found in the Fold: episode III. Building off their initial development and implementation during episode III, these Reshaped Agents feature updates to their AI algorithm, firmware, and software. This project represents continued investigations into their morphology, their performative behavior, and interactions in the physical world. The crux of the development in the update featured in the video above involve improvements to its ambulation - particularly its quarduped (four-legged) gait. This implementation is featured in Agent ά6 with Agents ά4 and ά5 behavior as a comparison. See video above. |
|
|
|
Reshaped Agent: ά6 (above) involves updates to camera-based object detection of its physical surroundings. The robotic agent attempts to find and categorize human beings but failing miserably. Modifications to the object detection algorithm is currently ongoing. Integrating ultrasonic sensing using high frequency detection of objects close to its body in order for dynamic object avoidance is planned. AI behavioral research assistance by Ryan Fernandez, Masters in Robotics and AI from Ira A. Fulton Schools of Engineering at Arizona State University. The acquisition of several MISBkits was supported by an ASU Herberger Research Building Investment (HIRBI) Grant funded in 2025.
|
|
Livestream detail of "Reshaped Agent: ά6", robotic performer with 3D printed |
|
Agent Marcgritte's neural model from simulated training in the virtual space is transferred onto Reshaped Agent: ά6. A synthetic training scenario example (above) shows the behavior of Agent Marcgritte which demonstrates it adaptively responding to obstacles within its environment. The goal of Reshaped Agent: Agent ά6 is similar with the original neural model improved upon with real-world behavior control variations in tandem with physical sensors. Getting up from falling as well as human and obstacle avoidance is currently in development. |
|