The Advantages of Olfactory Sensors

Image Source: molpix/Stock.adobe.com

By David Pike for Mouser Electronics

Published July 25, 2023

Sensors are crucial in modern technology, providing machines with the necessary information to function correctly. The latest generation of technology forms part of the Internet of Things (IoT), in which machines collect data about themselves and their environment and share it to create a network of autonomous devices. As the IoT depends on this shared information and feedback, sensors will become more critical than ever.

Designers have long sought to provide machines with equivalents to human senses. The human brain is superbly conditioned to understand the information that the sensory organs can collect. However, artificial sensors often need more sophistication. Early sensors could not interpret the information they gathered without the processing power that is now becoming available.

Many such devices, including light and proximity sensors, are further limited as they rely on a clear line of sight or, in some cases, physical contact to function correctly. With the ever-increasing level of sophistication in today's technology, especially the recent advances in practical artificial intelligence (AI) and machine learning (ML), designers can no longer rely on simple sensing technology.

Smell as a Machine Sense

The sense of smell, known as olfaction, is a form of chemical analysis of low concentrations of molecules that are suspended in the air. When these molecules encounter a receptor in the nose, the receptor transmits signals to the brain's parts responsible for smell identification. The sensitivity of the sense of smell depends on the concentration of receptors, which varies from species to species. For example, a dog’s nose is far more sensitive than that of a human, and dogs can identify concentrations of chemicals that are far too weak for humans to notice. However, detection dogs are also of limited value in an industrial environment.

Olfactory sensors offer versatile and unique advantages as a detection method. Unlike image recognition and other vision-based technologies, the scent does not rely on line of sight. Odors can be detected from obscured objects, allowing detection without the need for invasive procedures.

Using Smell in the Real World

The sense of smell is closely related to more traditional chemical analysis techniques, such as using a physical swab to test for narcotics or other substances. Using this procedure, a sample is collected using a sterile device that is then carried to an analysis machine. Not only is this time-consuming, but it also requires the operator to have access to the object they are sampling. In an airport application that depends on a free flow of passengers or cargo throughout the facility, this process creates a bottleneck familiar to any frequent traveler. As a result, security personnel must continue carrying out spot checks.

Olfaction does not require physical contact. Olfactory sensors can therefore be deployed unobtrusively around the security area, collecting information about passengers or bags. Equipped with a database containing the chemical signatures that cause concern, along with the processing power to analyze a very high number of samples in real time, these sensors would allow security personnel to let passengers flow through the facility easily, stopping only those who display traces of the chemical.

Many other applications could take advantage of this lack of physical contact. For example, some of these sensors are small enough to be mounted onto uncrewed aerial vehicles (UAVs). With onboard sensors that can detect chemicals in far lower concentrations than the human nose, UAVs can be programmed to traverse broad areas, searching for specific odors and reporting in real time using the latest 5G or satellite communications. From the early detection of wildfires to the identification of invasive species in agriculture, the usefulness of this technology is limited only by its database of chemical signatures and the sensors' sensitivities.

Olfactory sensors can also be partnered with other non-contact technologies, such as thermal imaging, to provide a suite of sensors. Using such a combination on a UAV could enable emergency services to respond rapidly to areas that have suffered a natural disaster. Survivors can be located quickly using this combination of techniques to provide the best possible chance of rescue.

Smell in the Industrial Environment

The industrial world is also adopting smell sensors. Many industrial processes have the potential to create hazardous byproducts. Olfactory sensors can monitor air conditions and highlight the dangerous buildup of harmful chemicals. They can also deliver meaningful information about the industrial process itself—for example, high concentrations of unburnt fuel in the atmosphere can result from incomplete combustion, a sign that a function uses energy inefficiently. Olfactory sensors, combined with the latest AI technology, can provide early warning of a problem and advise the best action to remedy the situation without human intervention.

This application even extends to monitoring the condition of the machine itself. For example, mechanical damage can create high temperatures that cause the loss of lubricating oils from the grease, leaving a dry, crusty soap that can cause the path to seize. By detecting the smell caused by changes within the chemical makeup of the lubricant, the problem can be detected before a failure can occur, allowing preventive maintenance to be scheduled.

One of the most exciting applications for olfactory sensors is in healthcare. Medical technology is constantly advancing; but despite the improvements in treatment technologies, early diagnosis is the most powerful tool for ensuring the best clinical outcome for patients.

Many conditions, from cancer to diabetes, detectably change the body's chemistry, and some dogs can even smell these changes long before any physical symptoms can be identified. Similarly, using sensors to detect changing odors can provide a critical early diagnosis that could significantly improve the chances of effective treatment and recovery. These sensors' non-contact, non-invasive nature further enhances their usefulness. They can be employed during an initial consultation without the delay associated with more traditional blood or tissue analysis techniques.

Techniques and Technology

How organic olfactory receptors, such as those found in the human nose, detect and identify smells has yet to be understood entirely. That said, we do know that the odor receptor nerve cells found in the nose are each equipped to identify one specific molecule, and developers are using different methods to replicate this functionality.

The exact process these sensors use to identify specific molecules varies between technologies. Some use organic semiconductors whose electrical characteristics change in predictable ways when exposed to specific molecules. Other techniques include mass spectrometry, which subjects a sample to a stream of electrons. The electrons ionize the sample into positively charged fragments, which then separate according to their mass-to-charge ratio. As each element behaves differently under these conditions, the results are displayed as a spectrum of characteristics that can identify the individual molecules present within.

Alternatively, some systems use ultra-fast gas chromatography for detection. In this method, the sample is dissolved in a fluid solvent, and its constituent parts separate at different speeds, creating a spectrum that shows the sample’s composition.

Advanced computing will enable the rapid analysis required to deliver results in real time. To create technology that can respond to the situations and abstraction of ordinary human activities, companies like Intel^® are seeking to emulate the neural structure and operation of the human brain. The Intel Loihi 2 chipset uses a technique called neuromorphic computing, in which the structure of the chip mirrors that of the human brain. This complexity is expected to provide the flexibility that AI needs to process the huge volume of information that these sensors will collect.

An alternative approach is being taken by Koniku, a biotechnology company using synthetic biology merged with silicon to build devices that can map common volatile organic compounds. Koniku refers to this technology as “smell cyborgs” that are capable of autonomously and intelligently reading the air. By engineering proteins in biological neurons to create precise protein-particle interactions, the biological cells can function as sensors, amplifiers, and biological signal processors. The hope is that this technology can be deployed as autonomous units to provide the functionality needed in a range of applications.

Regardless of the method used, the critical requirement for any smell sensor is its ability to accurately detect a range of molecules, such as those found in a complex organic compound. The combination of these molecules must then be analyzed and compared to identify the specific odor and allow conclusions to be drawn.

Conclusion

As a society, we increasingly depend on machines to help make complex decisions on our behalf. We must equip these machines with the required technology to collect the information needed to make these decisions. Sensors, therefore, play a crucial part in a vast array of applications.

Alongside traditional vision-based sensors, engineers are employing alternative methods such as olfactory sensors. When combined with advanced processing power, olfactory sensors deliver a range of advantages over other technologies. They require neither physical line of sight nor physical contact.

With applications in a wide range of industries and applications, from security and industry to ground-breaking medicine, olfactory sensors work in conjunction with other techniques to provide machine systems with the feedback they need to help improve lives.