Bio-Inspired Sensors and their disruptive potential

Pallawi
9 min readAug 18, 2020

Since my childhood, I have been among the curious audience of Wildlife channels and I am sure many of you must be. I have not just enjoyed watching them on television but have also been fortunate to be in the company of a few of them as pets or around them, animals like dogs, deer, fishes, cows, cranes, ducks, parrots, monkies, snakes, mongoose and a big garden with lots of flowers and trees collected from different parts of the country. As I am writing this blog I am thankful to my grandparents for they introduced me to this curious world of nature and taught me the fundamentals of the bond between human and nature. It is always fascinating to watch the activities of animals and plants and their supernatural abilities to perceive, adapt and respond to events so uniquely.

Today the fact claims that we humans are the most evolved species on earth. This is supported by countless experiments, inventions all focused to decrease the human mortality rate.

The facts say that we are ready for tomorrow but today and every day, we are hit by inexorable intruders. These intruders have not left any stones unturned to challenge our medical, chemical, military, geospatial, disaster management, agricultural, environmental science.

In this blog, my aim is to feed your curiosity and together explore the world of bio-inspired sensors. I am going to talk about how the world we live in uses them, can use them and could have used them for everyone’s better.

Biomimetics or Biomimicry: Innovation Inspired by Nature, is such a field that justifies how humans have looked at nature for answers to problems throughout our existence. One of the early examples of biomimicry was the study of birds to enable human flight. Although could not succeed in creating a “flying machine”, Leonardo da Vinci (1452–1519) was a keen observer of the anatomy and flight of birds and made numerous notes and sketches on his observations. The Wright Brothers, who succeeded in flying the first heavier-than-air aircraft in 1903, allegedly derived inspiration from observations of pigeons in flight.

Data — Demand, availability and Supply

Today every scientist is solving problem statement which spans around data. One common problem that we all face in this field is the requirement of precise and adequate data for our models. So why not pivot our research from gathering data using just one source. Why not identify and build data gathering architectures which are a fusion of data-focused bio-inspired sensors.

These architectures can deliver a precise, surplus, scarce and inaccessible data from our surroundings. I believe this can also shoulder the burden of computation cost and motivate ensemble model architecture.

Bio-inspired sensors solving human world data specific problems

1. Collection of data in low light conditions by nocturnals

Services provided by industries like sewage and mining are among the essential services to humans. Catering to such large consumers and working in unhygienic environment is challenging. Sewage pipes corrosion, no lights inside conditions, toxic pipe leakage, solid waste blockages, manual scavenging are few such problems that are commonly faced.

But the good news is that this industry is evolving not just by using cameras, infrared or ultrasonic sensors but also with the help of robots inspired by the Insects, Reptiles, Rodents and amphibians living inside the sewage and mining space. One such reptile which has changed how these spaces used to be monitored is Gecko through its eyes and feet.

Gecko Eyes and feet- The reptile Helmeted Gecko is nocturnal. They are among the very few living creatures able to see colours at night. Their multifocal optical system is comprised of large cones, which the researchers calculated to be more than 350 times more sensitive than human cone vision at the human colour vision threshold. The nocturnal geckos’ multifocal optical system is an advantage because the light of different ranges of wavelengths can focus simultaneously on the retina. Another possible advantage of their optical structure is that their eyes allow them to focus on objects at different distances. Therefore the multifocal eye would generate a sharp image for at least two different depths. Apart from their special eyes, Gecko feet are special, allowing it to climb steepest of surfaces. There are millions of hairs on their feet that stick using Van der Waals forces. Gecko feet architectures are used by NASA to stick to space objects.

2. New ways of Depth estimation by Salticidae Spiders

Depth estimation is an interestingly challenging field of study in computer vision. The depth image of a scene has information about the distance of the objects in the scene image from a viewpoint. Many approaches have been taken up to solve this problem and the most used among them are stereo imaging. Now with the emergence of deep learning, stereo image datasets are built and models are trained for depth estimation. But there are few living organisms whose eyes and its placements have opened new doors into this field, like Jumping Spiders.

Salticidae Jumping Spiders: Biological studies have revealed that jumping spiders have very unique and specialized optics, instead of a single layer retina as in the human being, they have multiple semi-transparent retina. They can capture multiple images of the same scene but the images captured on the different layers of the retina, have different levels of blurriness. They are able to extract the information in a very efficient way without using the stereo vision like humans do to estimate distance. They don’t even have the brainpower to process vision as we do. Harvard University developed a metalen depth sensor based on the comparison done at different levels of blurriness in the images captured by the multi-functional flat lense.

3. Odour detection and Chemoreceptors- Could be a Silver lining in detecting COVID-19, already helps to detect cancer and many others

A chemoreceptor is a molecular translater. It translates the information which is contained in the odour molecules like its structure, function, group, size into bits and bytes and these signals is then interpreted by brains.

How does the olfactory system in humans works — The odorant molecules are in the gas phase when outside the nose. These molecules need to sit on the nose receptors which are on the nerve endings in a space inside the nose also called the olfactory epithelium. Each receptor is sensitive to one particular type of molecule.

The olfactory epithelium is surrounded by an aqueous environment, which acts as a shuttle to carry the gas molecules and bind them with the receptors. This liquid space has proteins called G-proteins. Once the odour hits its designated receptor a cascade amplification is triggered as it binds to an ion channel and this allows the positive ions to move inside and cells to depolarize and fire an action potential. (In biology, depolarization is a change within a cell, during which the cell undergoes a shift in electric charge distribution, resulting in less negative charge inside the cell. Depolarization is essential to the function of many cells, communication between cells, and the overall physiology of an organism.)

Bunch of gaseous molecules are transmitted by these nerves and are classified into similar groups called the glomerulus. If the input gas was benzene then the benzene receptors in our nose would be activated and have a dedicated glomerulus for it in the olfactory bulb which will independently synapse (have a dedicated cell called mitral cell )with nerves that send the signal to our brain. A different set of molecules has different glomerulus and mitral cells.

But there is an animal which has outperformed the way Chemoreceptors work. The animal is Sniffer Dog which has 300 million receptors cells whereas we humans have just 5 million. Dogs have the ability to smell separately with each nostril. Smelling in the stereo helps to determine the direction of the source. Unlike our clumsy way of breathing in and out through same passage, Dogs exhale through the slits at the side of their nose creating a swirl of air that help draw in new odour molecule and allow odour concentration to build up over multiple sniffs. The olfactory system in dogs occupies more area as compared to humans allowing them to remember a staggering variety of specific scent at concentration 100 million times less than humans nose can detect.

Today Dogs are being trained to classify the smells from the sweat of a COVID-19 patient and that of a healthy person. Dogs can detect breast and prostate cancer by smelling the blood sample with 99% accuracy. They can give alerts for low blood sugar for the reason that they can detect Isoprene in the breath. Dogs can intimate you about the migraine that you might suffer due to their ability to sniff Serototine. They can alert you way before you diagnosed or suffer diseases like Malaria, Narcolepsy.

Several studies are going on to develop gas sensor materials inspired by dogs olfactory system.

5. Hydrophobic microscale bumps of Self-cleaning Lotus

Lotus plants stay dirt-free, and they do so without using any artificially created substance, unlike humans. The reason is Lotus cuticle is made up of soluble lipids embedded in a polyester matrix and the degree of its water repellency is extreme due to contact angle greater than 170 degrees which makes it superhydrophobic. The lotus leaf has a series of protrusions on the order of 10 μm high covering its leaf surfaces. Each protrusion is itself covered in bumps of a hydrophobic, waxy material that are roughly 100 nm in height. When water droplets are applied to the lotus leaf, they sit lightly on the tips of the hydrophobic protrusions as if on a bed of nails. This combined structure traps a layer of air between the surface of the leaf and the water droplet. Hence, the water is not allowed to wet the surface and is easily displaced.

The use of the bumpy architecture of their leaves has helped develop strain sensors. For example, when their hydrophobic coating is combined with a flexible silicon substrate, that could be stretched, bent or twisted, the coating can work as a wearable strain sensor that detects the changes in strain in real-time with response time in few milliseconds.

6. Early diagnosis and driving successful cancer surgeries by Mantis shrimp

Mantis shrimp: Optoelectronics focuses on light-emitting or light-detecting devices, and the pair have revealed that the mantis shrimp has the most complex ‘light detector’ in the animal kingdom. The mantis shrimp can see differences in polarized light — that is light that that is radiating in different planes of direction. It can detect both linearly and circularly polarised light. Polarised light refers to light that vibrates only in a single plane. Current research also shows that cancer cells do not reflect polarised light, while healthy cells do.

Researcher says that cancer cells are easy to see under polarized light because their disorganized and invasive structures scatter light differently (causes more depolarization) than normal body cells. Polarized imaging can spot cancer cells much earlier. The researchers are also experimenting with using polarized light to increase tissue contrast to help doctors tell where to start and stop cutting during surgery.

Conclusion:

I know we are busy identifying, solving and troubleshooting problems with the available resources. These inputs have been created using our existing technologies which is accessible to us easily. If we see the field of computer vision most of the inputs are derived using cameras and problem-specific lenses but there is a world which is parallelly developing Biomimetics. After reading and researching on many of such sensors inspired by nature I feel this is a valid ‘out of the box’ solution. Writing this blog helped me to stretch the horizon of my thoughts over the existence of sources which we can use to create data and encourage us to focus more on solving than troubleshooting.

References:

Gecko Lense: https://newatlas.com/geckos-next-gen-lenses/11661/

Gecko Feets:https://www.youtube.com/watch?v=neSyZDs79tE https://en.wikipedia.org/wiki/Gecko_feet

Last-mile delivery: https://www.youtube.com/watch?v=1FC12zhiTdY,

https://www.youtube.com/watch?v=0yMv16p8FO8

Polarization:https://www.youtube.com/watch?v=c9ew1J0PY-M,

Open source Arial dataset: https://lionbridge.ai/datasets/15-best-aerial-image-datasets-for-machine-learning/

Odour Sensors:https://www.youtube.com/watch?v=9ybGqwAyy3s,

Lotus self-cleaning:

https://onlinelibrary.wiley.com/doi/epdf/10.1002/adma.201702517, https://asknature.org/strategy/surface-allows-self-cleaning/,

https://www.youtube.com/watch?v=MU_tPRIrclE,

https://www.teachengineering.org/lessons/view/duk_surfacetensionunit_less4,

https://www.youtube.com/watch?v=PPJ0Khs7uWs

--

--

Pallawi

Computer Vision contributor. Lead Data Scientist @https://www.here.com/ Love Data Science.