Wilding Agents merges machinic behavior and plant community interactions to produce living, dynamic intelligences and resilient landscapes. The central proposition of this project is a behavioral rather than compositional model for design in which the material and logical relationship between agent and environment is codetermined: a rover catalyzes synthetic urban environments by automating ecological participation in undervalued urban space. The rover builds an intimate relationship with its environment as not its ward but its memory, as it reads and writes information onto the landscape through sensing environmental conditions and sowing seeds. This landscape inscription enables the vehicle to work on increasingly complex problems with very simple sensorimotor feedback, using a Braitenbergian model. The codependency built into this model puts machinic behavior in touch with plant community interactions.

The critical structural difference between ecosystems and contemporary urban infrastructures is that an ecosystem’s aim is resiliency, not stability. While urban infrastructure is built on a foundation of positivist frameworks of control, an adaptive design methodology can un-bind urban design approaches from deterministic urban fabric. Wilding Agents uses a multi-leveled behavioral model that accepts a range of hyper local inputs including field conditions, infrastructural capacity, and community preference. The model is comprised of a custom seed and plant behavior database; a community interface; and the rover, which uses light, moisture, temperature, and distance sensors to inform planting behaviors. Using a digital model to simulate rover and plant dynamics projects the model’s expansion into urban and ecological networks and connects it to larger-scale issues like stormwater management, habitat connectivity, and soil health.

Published ACADIA 2019. In collaboration with Nicolas Azel.