VIMOC’s Rosella™ IoT software stack is an end-to-end solution from the edge of the network to the cloud. This solution creates a seamless and robust link that allows intelligent infrastructure to drive intelligent applications.
Rosella is a distributed software solution performing sensory data capture and processing at the edge of the network, while efficiently cooperating with other computing nodes and Rosella’s cloud component. Each embedded software node performs data normalization, message parsing, edge processing pipeline and storage, with an extended framework for machine learning, vision processing etc.
Rosella’s cloud component communicates with the embedded computing nodes to authenticate and collect the intelligence produced by this network. This intelligence is presented via the Rosella API to enable advanced applications and services, including visualization, data mining/ analytics etc.
The Rosella platform will enable a new generation of applications to enhance customer convenience, lower labor costs, improve cash management and increase overall productivity. This is thanks to the Rosella machine learning framework, consisting of a library of deep learning machine vision algorithms for detection and classification tasks. The Rosella software can also connect to a variety of sensors that are commonly used in the ITS environment. This feature enriches the data available to Rosealla API developers to design new applications and enhance existing services.
The neuBox™ is an edge computing node. These nodes are capable of running Rosella™ and strengthen IoT infrastructure by absorbing the load of sensor data processing before passing it to the cloud.
Computing autonomy: If the link to the cloud goes down, the system continues to function, and refreshes data in the cloud when the connection is re-established.
Less traffic to the cloud, which results in more scalable LTE or M2M data use.
Lower latency compared to only cloud computing, allowing time-sensitive applications to be developed.
Leveraging existing infrastructure to enable new streams of data and new applications.
Pedestrian safety is of utmost importance in the design of shared vehicle/pedestrian zones. Solutions that improve visibility for drivers and pedestrians are key to reducing the risk of accidents. Pedestrians are the most vulnerable parties when such events occur, so VIMOC has implemented a rapid pedestrian detection system running on our Landscape Computing platform. We identify when a pedestrian is approaching a designated shared space such as a crosswalk and our system rapidly activates flashing LEDs embedded on that shared space to alert the driver of the collision risk ahead. The signal can also be sent to traffic signal control boxes as an input to either change the light’s state for traffic purposes or accident avoidance. The technology will also be key to ensuring the safe and efficient operation of autonomous vehicles, by strengthening their intelligence with intelligent and connected infrastructure.
Parking Garages offer urban environments a way to direct vehicles to a dedicated structure, resulting in a reduction of on-street congestion as drivers often precariously look for spaces close to their destinations. Garages form an integral part of cleaner and more pleasant urban spaces. The biggest hurdle to direct drivers to parking garages is the perceived risk of entering a full garage without the guarantee of a quick exit.
VIMOC has implemented a detection system using image sensors that leverages Landscape Computing’s powerful software architecture. Image sensors are positioned in view of entry/exit lanes and the feeds are sent to the neuBox installed on site. The neuBoxs process the live feed, applying advanced computer vision and classification algorithms which detect vehicles entering and exiting the garage, after which the raw images are discarded. The live results of these algorithms are available via VIMOC’s API, which can be used by any apps that are granted access to the data stream.
In the urban environment it is becoming increasingly important to understand multi-modal means of transportation. Pedestrian and bicycle transportation are two of the oldest transportation forms, and VIMOC has implemented a solution that enables a better understanding of how we use these forms of transportation in the 21st century. With this understanding, we will enable more intelligent, seamless and stress-free design and integration of infrastructure and enable new services and applications that leverage the rich and accurate information we are producing.
VIMOC has implemented a bicycle and pedestrian counting system using image sensors that leverage Landscape Computing’s powerful software architecture. Image sensors are positioned in view pedestrian zones and bicycle paths, and the feeds are sent to the neuBox installed on site. The neuBox processes the live feed, applying advanced computer vision and classification algorithms which detect bicycles and pedestrian numbers and direction of travel, after which the raw images are thrown out. The live results of these algorithms are sent to the cloud and available via VIMOC’s API, which can be used by any apps that are granted access to the data stream.