Building a relationship between the physical and software worlds. We use the Robot Operating System (ROS) in order to enable our machines to have their kinematics, perception, and control. We utilize deep learning techniques to improve computer vision and control tasks with the very same techniques used for adaptive cruise control. We then go further with our experience in evaluation & validation, supervised learning, unsupervised learning, and reinforcement learning to fine tune your specific application whether it is a prototype in Keras or a custom TensorFlow application.
Our custom machines are first simulated in software, like Gazebo, before testing and implementation. To aide in safety we utilize sensor fusion, going as far as to implement Unscented Kalman Filters (UKF) rather than just measurements from vanilla extended kalman filters. Our team are experienced with various technologies from low level power electronics all the way up to high level software to provide you with a truly custom service. These machines are built to meet ISO certification standards and various international wireless communication standards. The safety of your people is of the highest priority and will outweigh the requirements of automation.