MORE LOAD.
LESS ROAD
REAL-WORLD ROUTING
The Route Engine does not calculate tours from idealized models – it learns from real driving behavior. It analyzes how drivers use streets, how they navigate tight access points, and how routes change depending on the time of day. Machine-learning models process these patterns and turn them into routes that remain stable in real operations and reduce uncertainty.
Tech Stack
-
PostGIS for spatial analysis, bottleneck detection, road networks, and site evaluation
-
OSRM or Valhalla for routing graphs, path calculation, cost profiles, and fast route retrieval
-
Mapbox for real-time visualization of roads, routes, access points, and maps
-
Python for routing logic, data processing, and interface control
-
BINAI Driving Pattern Models for behavior analysis
PREDICTIVE ROUTE INTELLIGNECE
The Route Engine uses AI to predict how fill levels will develop, how road segments will be loaded, and which routes will remain efficient.
It anticipates future tours and automatically adapts to changing conditions.
Tech Stack
-
TensorFlow for time-series models forecasting fill levels, traffic behavior, and route evolution
-
Scikit-learn for classification, pattern recognition, and load forecasting
-
PostgreSQL for structured storage of historical and operational data
-
Python for model logic, data pipelines, and predictive calculations
ADAPTIVE LEARNING MODEL
With every tour driven, the model becomes more precise.
The Engine processes GPS traces, stops, reverse maneuvers, and micro-decisions to learn how real workflows emerge. Models continuously update themselves so that routing decisions align more closely with real-world patterns over time.
Tech Stack
-
Python Behavior Models for analyzing GPS, speed, stops, and driving dynamics
-
TensorFlow for continuous training and retraining of learning models
-
Scikit-learn for feature engineering and model refinement
-
PostgreSQL / PostGIS for storing and analyzing movement and location data
The Route Engine combines human experience with machine precision, creating a routing intelligence that sharpens with every movement.
It captures more than data points – it captures real behavior, in-the-moment decisions, everyday routines, and subtle patterns that only become visible through daily practice.
These signals shape models that understand how routes feel, how streets react, and how waste flows behave in real environments.
This creates routing that doesn’t think abstractly – it responds concretely.
It recognizes paths that bring stability, situations that slow operations, and factors that make a tour efficient.
It is a system that does not rely on rigid rules, but becomes more precise with every kilometer driven and continuously adapts to real-world conditions.
Based on this understanding, the functions of the Route Engine unfold.
They use learned patterns, turn movement into performance and transform experience into measurable efficiency.
SCENARIO AND SIMULATION
The Route Engine simulates how territories evolve, how volumes rise, or how fleets need to be adjusted. The best configuration can be tested long before it enters real operation.
Tech Stack
-
Python Simulation Engine for scenario calculation, variants and routing strategies
-
PostgreSQL / PostGIS as the foundation for territory models, street profiles and site networks
-
TensorFlow / Scikit-learn for analyzing and comparing scenario models
-
Geospatial Queries for evaluating topography, density, service zones and traffic patterns
LOAD AND EFFICIENCY OPTIMIZATION
Die Route Engine optimiert Nutzlast, Kilometer, Energieverbrauch und Zeit gemeinsam. Sie erkennt Schwachpunkte im System und priorisiert Wege, die die Gesamtleistung steigern.
Tech Stack
-
Python Optimization Framework for multi-dimensional optimization of time, load and energy
-
Scikit-learn for clustering and pattern analysis for routing decisions
-
PostGIS for evaluating spatial parameters and bottlenecks
-
Performance Models in TensorFlow for calculating and prioritizing efficient route variants
Why it matters
Routing determines cost, effort, time and CO₂.
Every unnecessary trip, every traffic jam and every poorly placed container creates pressure on the entire system.
The Route Engine addresses exactly this point.
It turns intuition into predictability, experience into a learning model, and daily tours into a foundation for better decisions.
The result is less idle time, higher payloads and processes that feel noticeably more stable and efficient.
Part of the BINAI System
The Route Engine never operates in isolation.
It uses signals from Intelligence, transforms them into concrete routes, and makes results visible through the Dashboard while sending updates and alerts back through Communication.
Data becomes decisions, decisions become movement, and movement becomes new learning signals.
This closes the loop – and makes the entire system better with every single tour.
Outcomes
Fewer kilometers per ton
Higher payload per tour
Less operational stress
More predictability across the system
And a data foundation that makes every decision traceable
Real Routes.
Trained by Humans