Predicting nonlinear control with Transformers, high-precision recognition through segmentation.
At Neorium Technology, we support model-based development using control technology and machine learning. This content introduces cases involving the construction of predictive models for nonlinear control targets using deep learning models such as "Transformer" and "LSTM," as well as high-precision image recognition (semantic segmentation). We can predict future behaviors of control targets with time delays and complex nonlinearities through time series learning, even handling phenomena that are difficult to model. Additionally, in image recognition, we accurately identify targets even in complex backgrounds using segmentation with Mask2Former, making it applicable in industrial and autonomous driving fields. We also support integration with control system design and experimental data analysis based on Python + TensorFlow and PyTorch. If you are interested, please contact us through our website. https://www.neorium.co.jp/contact-us/
Inquire About This Product
basic information
Deep Learning, Transformer, LSTM, Mask2Former Application 1: Prediction of nonlinear control targets (e.g., vehicle speed prediction) Input: Time-series sensor data from the vehicle (gradient, throttle, torque, regenerative braking) Prediction target: Vehicle speed 1 to 5 seconds ahead Application 2: Image recognition through semantic segmentation Input: Image data from urban environments (ADE20K) Prediction target: Segmentation of objects in the image (buildings, roads, signs, etc.) Development environment: Python + TensorFlow / Keras, MATLAB/Simulink
Price range
Delivery Time
Applications/Examples of results
This technology has a proven track record in building vehicle speed prediction models, capable of accurately predicting vehicle speed a few seconds ahead based on time-series data such as regenerative braking, throttle opening, and gradient. Additionally, in image segmentation, it has experience utilizing datasets like "ADE20K," which includes urban landscapes, to build multi-category recognition systems. Applications include advanced driver assistance system (ADAS) development, in-factory image inspection, autonomous mobile control, and robot vision, making it widely applicable in fields that require next-generation control and recognition using AI.
catalog（1）
Download All Catalogs
Company information
Neorium Technology responds to the diverse needs of customers involved in model-based development, consistently providing high-quality technology. In particular, we have a wealth of experience in professional services (model development, contracting, consulting, technical seminars) and product services (sales of control design support software and technical support) in the fields of control, optimization, machine learning, and system modeling. If you have any issues with model-based development, please feel free to consult us. Additionally, in collaboration with our partners, we also offer support services for ADAS simulation testing in the automotive sector. Examples include road surface creation, creation of OpenDRIVE and OpenSCENARIO data, and automation through MILS/HILS.