Benchmarking and enhancing the performance of insect-based chemical sensing biorobots in static and dynamic environments
Abstract
Biological systems have optimized their transducers and sensing systems over evolutionary timescales to achieve capabilities that cannot be matched by state-of-art engineering systems. This is particularly true for chemical sensing, which is the most primitive of the sensing systems, yet the least well-understood. While reverse-engineering the brain and biological sensory systems remain the holy-grail in understanding complex systems, can forward-engineering approaches be developed that directly exploit the sensing capabilities of even relatively simple invertebrate models to solve practical engineering problems? An approach that part integrates biological components, especially the chemical sensors and their sophisticated processing machinery, with engineered modules that augments their capabilities and allow precise control would synthesize the best of what both artificial and natural systems offer. Our current work has shown that explosive vapors from a variety of chemicals including TNT, DNT, RDX, PETN and ammonium nitrate evoke neural responses in the locust brain that are unique and therefore can serve as a fingerprint for their recognition. However, systematical benchmarking of its performance, and enhancing current capabilities achieved remain to be carried out. Towards this goal, we propose the following main aims: Aim 1: Benchmark the explosive sensing performance of the insect based bio-robots: While our prior work establishes the feasibility of an hybrid chemical sensing approach that takes advantage of the biological chemical sensing array and subsequent neural signal processing framework, there are several questions that still remain to be answered: • What are the limits of detection (LOD) for these explosive vapors? • How sensitive are the neural responses to small changes in the encountered concentration of these explosive vapors? Such sensitivity would be essential for localizing the odor source using concentration gradients. • Are the responses evoked at different intensities still consistent to allow robust recognition of the odorant in an intensity-invariant fashion? Aim 2: Enhance the explosive sensing performance of the insect based bio-robots: The overall objective of this aim is to develop methods that will allow fast, rapid and random neural implantation, yet allow robust chemical sensing with insect based biorobots. Such approaches would be necessary for moving this technology from ‘laboratory only’ setups to rapidly deployable, and field ready units. Aim 3: Integrate methods to develop a ‘plug-and -play’ style chemical sensing system and systematically characterize performances in a wind tunnel: We will also develop next generation of adaptive, miniaturized instrumentation that interfaces and taps the signals readout with these electrode arrays. The goal is to develop a ‘plug-and-play’ style system with significantly reduced prep time (currently more than hour to ten minutes or less) and the amount of training needed to create these insect based chemical sensing system. Finally, we will quantify the performance of this enhanced preparation in a wind tunnel where richer odor plume dynamics mimicking more realistic scenarios will be explored. In sum, the overall goal is to fully benchmark the insect based explosive sensing approach, compare its capabilities with other model organisms, and enhance the overall chemical recognition performance to take it several steps closer towards development of real world, field deployable units.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 06, 2021
- Source ID
- N000142112343
Entities
People
- Barani Raman
Organizations
- Office of Naval Research
- United States Navy
- Washington University in St. Louis