Abstract:Time-of-Flight (ToF) cameras are becoming popular in a wide span of areas ranging from consumer-grade electronic devices to safety-critical industrial robots. This is mainly due to their high frame rate, relative good precision and the lowered costs. Although ToF cameras are in continuous development, especially pulse-based variants, they still face different problems, including spurious noise over the points or multipath inference (MPI). The latter can cause deformed surfaces to manifest themselves on curved surfaces instead of planar ones, making standard spatial data preprocessing, such as plane extraction, difficult. In this paper, we focus on the MPI reduction problem using Feature Pyramid Networks (FPN) which allow the mitigation of this type of artifact for pulse-based ToF cameras. With our end-to-end network, we managed to attenuate the MPI effect on planar surfaces using a learning-based method on real ToF data. Both the custom dataset used for our model training as well as the code is available on the author's Github homepage.
Abstract:Global positioning systems can provide sufficient positioning accuracy for large scale robotic tasks in open environments. However, in underwater environments, these systems cannot be directly used, and measuring the position of underwater robots becomes more difficult. In this paper we first evaluate the performance of existing pose estimation techniques for an underwater robot equipped with commonly used sensors for underwater control and pose estimation, in a simulated environment. In our case these sensors are inertial measurement units, Doppler velocity log sensors, and ultra-short baseline sensors. Secondly, for situations in which underwater estimation suffers from drift, we investigate the benefit of intermittently correcting the position using a high-precision surface-based sensor, such as regular GPS or an assisting unmanned aerial vehicle that tracks the underwater robot from above using a camera.