Hidden Markov models applied to the validation of sensor transient responses

Methods for using the transient response of composite polymer chemiresistors for validating sensor behavior are presented. The transient responses of ten types of (composite polymer) chemiresistors are characterized using hidden Markov models and then tested. Results show that relatively small hidden Markov models (four states or less) can be used to accurately distinguish the transient responses of closely related chemiresistors. These models can be used as a means to establish a confidence level with which subsequent interpretations of sensor response (individual or array based) including analyte discrimination can be made.