(less than 20 mJ per second). the issue of continuously collected sensory readings overloading the storage in the smartphone can become problematic. The core challenge here is how to design techniques that enables the smartphone continuously sense the ambient environment for vibrations. but retain only those readings within a time window that correspond to an explosion, while filtering out the rest. We have done prior work in {11} to make this goal feasible. The rationale of our technique is to identify appropriate thresholds for the ratio of sudden spikes in vibration during an explosion to the long term dormant vibration readings in the absence of an explosion. Leveraging our findings in [I l], we implemented an algorithm and installed it as an app on the smartphones to retain only those accelerometer readings corresponding to an explosion, while filtering out the rest. This dramatically improves storage efficiency. We point out that our app consumed 331KB of memory, and power consumed was 0.173 joules per second, both of which are quite minimal