Manini Banerjee

   

Systems Designer & Research Engineer making environmental complexity legible and actionable. 



COMPUTATIONAL ECOLOGY: 
BIOPOD Co.
ECOLOGY · INFRASTRUCTURE · SYSTEMS
Designing deployable ecological infrastructure for wetland restoration based on environmental research.
Ecological AI 
PREDICTION · INTERFACE · DATA  
 A Decision-Support System for Targeted Ecosystem Restoration.
Algorithmic Morphogenesis
BIO-COMPUTATION · DATA MATERIALIZATION
inscribing real-time human neurological data (EEG) into living algal morphology using phototactic actuation


HARDWARE & INTERFACES: 
Threads

EDGE ML  ·  HARDWARE  ·  TEXTILES
Sentient Surfaces + Edge ML on Textiles. Human-AI Symbiosis through Ubiquitous Computing.
S(kin)-orb
HAPTICS · BIOSENSING · AFFECTIVE COMP.  A bio-digital interface translating electromyographic (EMG) signals into haptic feedback for remote affective communication.
Vermiform

COMPONENT · SOFT ROBOTICS · WEARABLE
Bio-mimetic architectures for wearable computing. 

Chito-bot
BIOCOMPOSITES · TRANSIENT ELECTRONICS 
Investigating material compliance and structural integrity in bio-composite hexapods.



STRATEGIC SYSTEMS: 
PFV

MOBILITY  ·  ECOLOGY  ·  SYSTEMS

Autonomous Mobility as Urban Bio-Infrastructure.
Aero
SENSING · MATERIALS · DATA  
Developing robotic material systems for localized air-quality sensing and pollutant sequestration through embedded environmental intelligence.
Bio - intelligence
COPMUTATION · SYSTEMS ·  BIOLOGY 
Exploring biological computation as an alternative model for system intelligence and control.


Archive 

© 2019-2026 Manini Banerjee

Bio-intelligence

bi·o in·tel·li·gence

Designing cars powered by organisms, not algorithms



Details
Microbiological Culture • Robotics Design • Technological Anthropology

Client
Hyundai Motor Group ~ Hyundai Future Experience Research Collaborative 2022

Date
2022

Team
Mehek Vohra • Manini Banerjee • Paula Gaetano Adi

Role
 Concept, Microbiological Culture

As seen in:
"Driverless Cars' Need for Data is sparking a new Space Race." Autonomous vehicles will generate as much as 40 terabytes of data an hour from cameras, radar, and other sensors—equivalent to an iPhone's use over 3,000 years.... (Bloomberg)  


Data processing, storage, and production require carbon-emitting data centers.  Through 
Bio-intelligence, an alternative paradigm to Artificial Intelligence, presents an opportunity for sustainable, data-free navigation.  

How would a car work if organisms instead of algorithms drove it?  

What if we moved from nature-inspired to nature-collaborated?  

For many years, we rode horse carriages and controlled the horse rather than allowing it to operate using its intelligence. This freedom is the essence of Bio-intelligence.




We researched the fungal animal and living memristor: Slime Mold, Mycetoza, Physerium Polycephalum, or
"The blob" during this project. 

We chose this organism as it:

1) Uses Phototaxis and Chemotaxis instead of sensors, cameras, and data.
2) Does not rely on electricity for data storage, and memorizes past routes.
3) Operates as a collective and learns past routes.
4) Demonstrates generalizability in navigating new + old terrains

Slime mold demonstrates intelligence - it can grow, learn, predict, and adapt depending on its environment. It can remember where it has previously been, and can navigate new terrains without having to be manually trained.




The mould shoots out protoplasmic tubes searching for an efficient path toward oat flakes.

In one of the experiments we conducted, we placed its favorite food - oats - at specific locations in a petri dish and recorded its growth over 3 days. It ended up creating efficient paths between each oat, proving its navigational intelligence.

Autonomous vehicles require extensive training to memorize and learn routes. Is there a way slime mould can collaborate with these vehicles to make them more efficient?
We set up an experiment to test our theory.


1) Robot navigation: Has to navigate through every section of the maze before finding the most efficient path. This process requires a significant amount of time, computation, and data storage.


2) Slime Navigation: The same maze being solved by the Slime mold - uses no sensors, processing algorithms, or data storage.
Lorem Ipsum...
Lorem Ipsum...
However, whilst working with living matter, there are challenges regarding the speed of growth and life spans.

Is there a way we can immortalize Biointelligence?

To tackle such questions, we created a grasshopper simulation to mimic slime growth. 

In the simulation, we would input the start and end points and allow the simulated Slime Mold to calculate the most effective path. This path is then fed into the robot.

The resolved form is a camera mounted over the autonomous robot carrying a smart petri navigating cassette. We imagine computer vision being used to read the slime mould's movements and inform the vehicle's navigation. The smart petri contains self-healing agar and programmable oats to simulate a range of paths by harnessing the Biointelligence of the organism.