System Energy Efficiency Lab
Home People Research Publications Sponsors Contacts
<- back

SHiMmer software

System Tasks

SHiMmer is designed to perform active sensing, processing, and communication. Operating system primitives such as interrupt service routines, radio communication, and energy management are handled by the microcontroller. The DSP handles the high-level aspects of structural health monitoring, such as the wave generation, response sampling and data processing. An example flow of SHiMmer's tasks are shown in the figure. The DSP actuates and samples a wave using piezoelectric transducers (PZTs), where one path is defined as the pairing of one PZT as an actuator with another PZT as a sensor. The DSP then averages the data from multiple iterations of a single path, followed by filtering using a bandpass or matching filter. After this, the data is divided into blocks where a feature is calculated for each block. The features from many paths can then be correlated for a specific point of interest to determine if damage exists. When processing is complete, the microcontroller transmits the results to the external device and waits for further instructions (likely, returning to sleep mode.) Additionally, the microcontroller performs energy management, described in further detail below.

Energy Management

Because SHiMmer is an energy harvesting system (obtains power from the environment - solar power), it can operate without human intervention for long periods of time. As a result, however, a significant challenge is the management and conservation of energy. SHiMmer must maintain energy neutrality, meaning that it can only consume as much energy as it can harvest. Thus, the overall goal of SHiMmer is to execute a sequence of tasks, such as those presented in the figure, while maintaining energy neutrality.

Depending on the time of day and the amount of stored energy, SHiMmer may have to wait to complete all the tasks until additional energy can be harvested. To determine how long to wait, SHiMmer employs a predictor and a controller, as shown in the diagram to the right. The predictor maintains past energy harvesting information to predict future harvesting rates, while the controller determines which tasks to execute and when to execute them.

The proposed system manager adapts the task allocation and their complexity over different days based on the amount of energy collected by the on-board Solar Energy Harvester (SEH). As a result, depending on the weather condition, e.g. sunny or cloudy day, the system modifies the frequency SHM measurement accordingly.