Revolutionizing East Mining Company's Road Management

Problem: The existing system was functional, yet lacked some vital features

East Mining harnesses coal through the open-pit technique. Unlike subterranean methods, this approach involves excavating directly from the terrain's surface to access and extract coal layers. Consequently, the landscapes of their operations are dominated by vast, open-air, funnel-like depressions intertwined with a dynamic network of roads.

Operations persistently unfold across diverse regions, implying that pathways in every zone undergo alterations daily. These passages are navigated by specialized dump trucks, such as the BelAZ models, which overshadow standard dump trucks in size and can transport cargo volumes 3-10 times greater. This increased capacity, however, leads to augmented fuel consumption, necessitating that their travel paths be optimized to minimize halts and avoid congestion.

Data pertaining to vehicular activity is meticulously recorded at intervals of a mere second. This granularity aids dispatchers in visualizing vehicular flows and pinpointing possible congested zones. Armed with this insight, they can re-route vehicles or conceive alternative pathways to alleviate busy corridors. Such adaptability not only curtails fuel and workforce expenditures but also considerably amplifies the rate and overall quantum of coal moved within a given timeframe.

Though East Mining had previously instituted a dedicated system for route oversight, it lacked comprehensive features. A key limitation was its perception of roads as monolithic entities, whereas a more granular segmentation was sought. Recognizing these gaps, the Singula team stepped in, aiming to architect a more intuitive solution that could accentuate operational efficacy and returns.

Solution: Creating a state-of-the-art road management service

In a span of three months, Evrone and Singula partnered to devise a new road visualization service for East Mining. We initiated by liaising with the client's technical team to discern requirements. Essential features added included map editing, road management, versioning, and access control.

Frontend

We identified an open-source mapping library familiar to the operators. This was complemented with a JS library for element editing, making it akin to a graphic design tool. Data from machinery was represented as points on the map, with numbers potentially reaching over 10,000.

Backend

To illustrate the pathways that vehicles traverse on the digital interface, the backend interprets data, refines the spatial coordinates and timestamps, and relays the transformed data to the frontend.

Among the challenges we confronted was the intricate handling of data. The server's responsibility wasn't confined to merely acquiring telemetry and transmitting it to the front end. It also had to authenticate modifications introduced by the frontend. For instance, if an operator sketches a new roadway on the interface, it's the backend's duty to catalog this as an updated version of the map.

To address this, we engineered an API that outlines the series of alterations on the map encompassing roads and polygons. This API orchestrates the essential tasks to modify, introduce, or eliminate segments. To ensure these modifications were accurate, we implemented a structured rule set governing the interplay of different elements. This Rest API was constructed using django-rest, a choice that significantly accelerated the development phase and streamlined ongoing maintenance, given the platform's extensive built-in capabilities.

For API documentation, we opted for Swagger. This choice proved invaluable due to its semi-automated nature. Whenever a new endpoint materializes in the backend, its details instantaneously get reflected in Swagger, facilitating seamless communication with the frontend and efficient debugging. This consolidation of backend details in one location proves immensely beneficial for developers.

DevOps

Capitalizing on Kubernetes and full-fledged CI/CD, we opted for straightforward deployment to a virtual machine. This ensured swift operations from testing to deployment. Using a mono-repository approach, separate pipelines were built for the backend and frontend, expediting the build time.

Result and future prospects

East Mining now boasts a sophisticated service that dynamically updates the condition of their roads and dumps. Users can visualize, edit, and save road data, with error-checks in place to ensure data integrity. They can also toggle between past and present maps, and access is bifurcated into 'editor' and 'read-only' roles for increased security.

The segmented road data furnishes East Mining with precise analytics, aiding them in making data-driven decisions to enhance operational efficiency and reduce costs.

Future iterations might see automation features like AI-driven road mapping, contingent on accumulating enough data from human operators.

Post-project, East Mining took the reins to augment the system's functionality. They integrated the service into their OES IoT platform, employing the data to oversee road conditions, guide dump truck drivers, balance truckloads, and accomplish other business operations.