Sensor Fusion and Object Tracking using an Extended Kalman Filter Algorithm — Part 1

<p>This algorithm is a recursive two-step process:&nbsp;<em>prediction</em>, and&nbsp;<em>update</em>. The&nbsp;<strong>prediction step</strong>&nbsp;produces estimates of the current variables along with their uncertainties. These estimates are based on the assumed model of how the estimates change over time. The<strong>&nbsp;update step</strong>&nbsp;is done when the next measurements&nbsp;<em>(subject to noise)</em>&nbsp;is observed. In this step, the estimates&nbsp;<em>(let&rsquo;s call it&nbsp;</em><strong><em>state</em></strong><em>&nbsp;from here on)&nbsp;</em>are updated based on the weighted average of the predicted state and the state based on the current measurement. A lower weight is given to that with a higher uncertainty.</p> <p><a href="https://medium.com/@mithi/object-tracking-and-fusing-sensor-measurements-using-the-extended-kalman-filter-algorithm-part-1-f2158ef1e4f0"><strong>Website</strong></a></p>
Tags: Sensor Fusion