Self-Calibration of Accelerometer Arrays (SCARS)

惯性测量单元（IMU）以物体的横向加速度和角速度的形式确定物体的相对运动。常规地，IMU包括三个加速度计和三个陀螺仪。与此相反，无陀螺仪惯性测量单元（GF-IMU）仅基于加速度测量来检测相对运动。因此，附接在身体内的不同位置处的多个换能器共同形成加速度计阵列。为了准确地估计相对运动，传感器姿态，即，它们的位置和方向必须精确地知道。然而，这些往往只能在准确度不够的情况下获得。因此，它们通常通过校准来重建。现有技术的校准方法基于一组运动数据和对应的加速度测量来确定几何传感器配置。因此，需要复杂的实验室设备来对传感器阵列施加参考运动，这导致每个生产的测量单元的额外成本。此外，该设备大多不适用于具有大几何尺寸的传感器阵列，这抑制了许多应用的校准。本研究的目的是建立一种加速度计阵列的自校准方法。因此，加速度计姿态仅使用它们自己的测量来确定，而没有任何外部参考。此外，推导出的方法旨在对任意运动进行工作，使得其可以在GF-IMU的正常操作期间应用。为了实现这些目标，我们将研究如何引入固有的知识测量系统的传感器姿态的数值优化。由此产生的方法不是特定的加速度计阵列，因此有可能被其他测量系统采用。

An inertial measurement unit (IMU) determines the relative motion of a body in the form of its transversal acceleration and its angular velocity. Conventionally, an IMU comprises three accelerometers and three gyroscopes. In contrast to this, a gyroscope-free inertial measurement unit (GF-IMU) detects the relative motion based on acceleration measurements only. Thereby, multiple transducers attached at distinct locations within the body jointly form an accelerometer array. To accurately estimate the relative motion, the sensor poses, i.e., their positions and orientations, must be known precisely. However, those are often only available with an insufficient accuracy. Hence, they are commonly reconstructed by calibration. Current state of the art calibration methods determine the geometrical sensor configuration based on a set of motion data and corresponding acceleration measurements. Sophisticated laboratory equipment is thus required to impose a reference motion on the sensor array, which results in additional costs for each produced measurement unit. Moreover, the equipment is mostly not suitable for sensor arrays with large geometrical dimensions, which inhibits a calibration for many applications. The goal of this research project is to create a self-calibration method for accelerometer arrays. Thus, the accelerometer poses are determined using only their own measurements without any external reference. Moreover, the derived method is intended to work on an arbitrary motion, such that it can be applied during the normal operation of the GF-IMU. To achieve these goals we will investigate how to introduce inherent knowledge about the measurement system to a numerical optimization of the sensor poses. The resulting methodology is not specific to accelerometer arrays and thus has the potential to be adopted for other measurement systems.