The Electronic nose (EN) system based on an array of gas sensors and its recognition techniques have been widely investigated in recent decades. However, the integration of an EN system is yet to be demonstrated. The pre-processing and gas recognition algorithms of the reported EN systems are usually very complex and are realized by software in personal computers (PCs). This results in complex systems, that are typically expensive and power hungry. Human olfactory system shows powerful and efficient gas discrimination capabilities. However, little work has been done on such bio-inspired algorithms....[ Read more ]

The Electronic nose (EN) system based on an array of gas sensors and its recognition techniques have been widely investigated in recent decades. However, the integration of an EN system is yet to be demonstrated. The pre-processing and gas recognition algorithms of the reported EN systems are usually very complex and are realized by software in personal computers (PCs). This results in complex systems, that are typically expensive and power hungry. Human olfactory system shows powerful and efficient gas discrimination capabilities. However, little work has been done on such bio-inspired algorithms.

This thesis presents an integrated EN system based on an array of gas sensor fabricated in our in-house foundry. A thorough characterization of the sensor is performed and a novel parameter extraction technique is reported. Moreover, two different spike-timing encoding schemes are developed and verified. These schemes encode the sensor's responses into spike trains which are unique for specific gases and are invariant to gas concentrations. A prototype is built to realize the gas detection using our sensor array and the proposed encoding schemes. This prototype is a compact, low power and low cost EN system, which integrates the data acquisition, signal preprocessing and gas recognition. A novel voltage to time conversion scheme is proposed enabling to remove the need for an Analog-to-Digital Converter in our integrated system. The experimental results on real gases show an excellent real-time performance in terms of gas recognition ability.