Perception for Autonomy
While GNSS provides precise, real-time positioning information to answer the question, “Where am I?”, perception provides information about the surrounding environment to answer, “What is around me?” The ultimate goal of perception is obstacle detection so as to avoid people or other vehicles, or navigate in relation to reference objects, such as crop rows in the case of agriculture, whilst maintaining a high level of productivity of the machine.
There are several types of sensors that can be used individually or in concert to achieve this goal, including high resolution mono and stereo cameras, radar and LiDAR. Of course, sensors only provide the raw information. Machine learning — the algorithms behind autonomy — is required to “teach” the perception system what’s what for object classification as well as the operating environment. Hexagon | NovAtel have been developing these algorithms based on the latest sensor and machine learning technologies to allow our customers high levels of performance and robustness expected within Agriculture.
Combining the positioning, sensor and platform expertise of Hexagon, we are investing in the development of machine perception by testing, validating and refining both the hardware and software elements of perception systems to accelerate off-road autonomy. There’s a lot to consider when choosing the best options for your autonomous system, and we can help you navigate, whether you’re looking for advice on individual sensor types or full perception system development.