Effect of Sensor Error on the Assessment of Seismic Building Damage

Abstract

Natural disasters affect structural health of buildings, thus directly impacting public safety. Continuous structural monitoring can be achieved by deploying an internet of things (IoT) network of distributed sensors in buildings to capture floor movement. These sensors can be used to compute the displacements of each floor, which can then be employed to assess building damage after a seismic event. The peak relative floor displacement is computed, which is directly related to damage level according to government standards. With this information, the building inventory can be classified into immediate occupancy (IO), life safety (LS) or collapse prevention (CP) categories. In this work, we propose a zero velocity update (ZUPT) technique to minimize displacement estimation error. Theoretical derivation and experimental validation are presented. In addition, we investigate modeling sensor error and interstory drift ratio (IDR) distribution. Moreover, we discuss the impact of sensor error on the achieved building classification accuracy.

Accuracy Limits of Embedded Smart Device Accelerometer Sensors

“…We propose an error model that takes into account sampling time uncertainty as well as additive noise. Since, the additive noise parameters can be well estimated using power spectral density or Allan variance techniques illustrated in [20], [21], in this work we focus on estimating the sampling time uncertainty standard deviation (STD). In other words, we assume additive noise parameters are known and given.…”

“…Most prior work modeled the non-deterministic inertial sensor error as additive noise and characterised it using either power spectral density or Allan variance techniques as mentioned in [20], [21]. Sampling time uncertainty was neglected due to the fact that using a stable clock and proper sampling, results in negligible sampling time jitter error.…”

“…σ w and σ z . However, since σ w can be estimated while the device is at rest as described in [20], [21], hence, we assume σ w is known.…”

“…Let the sampling clock have jitter of Gaussian distribution with zero mean and σ z = 0.2. For smartphone accelerometers, according to [20], [27], the additive noise typically has σ w = 10 mm/s 2 .…”

“…Inertial sensors error is modeled in previous work [18], [19] as: 1) Scale factor error, 2) constant bias, 3) nonlinearity and 4) additive random noise. While the first three error types are deterministic and can be compensated by calibration, the random noise cannot be calibrated and can only be characterized as mentioned in [20], [21]. However, using the above traditional error model for a smart device accelerometer (or other sensors as well) does not capture the operating system (OS) timing uncertainty.…”