Miniature robots present a number of challenging problems in controls, as they often exhibit nonlinear dynamics and have strict power and size constraints. These constraints limit the sensing and processing capabilities drastically. Many control techniques require knowledge of the robot’s position, so the position must be estimated when it cannot be sensed directly. We report a mixed signal odometry circuit that maps motor commands to estimated and predicted changes in position in Euclidean space (x, y, θ). We compare the mixed-signal implementation with other approaches and find that the mixed-signal implementation offers significant reductions in power consumption at an acceptable loss of precision.