Technology Cornerstones
Combain offers cutting-edge positioning solutions designed to accurately track any connected device, whether indoors in a building or outdoors. The sophisticated algorithms behind these solutions, honed through more than two decades of dedicated research, ensure exceptional accuracy that meets the demands of various applications. While the positioning technology is advanced, integrating the Combain Location API into existing systems is seamless and straightforward. Combain’s innovative positioning technology is deployed in various industries across the globe.
Locate Everything Everywhere
Our cutting-edge location services, powered by AI and supported by an extensive crowdsourced location database, ensure unparalleled accuracy, reliability, and scalability. This equips our clients with the tools necessary to enhance user experiences and streamline operations across various industries, ultimately adding value to their businesses.
- Seamless outdoor / indoor
- AI-enhanced
- Crowdsourced location data
- Combain cloud or on-premises deployment
Cloud-based positioning solver
Wireless-based positioning Technology
It is all about radio environment and signal strengths
Location tracking has made significant advancements over the past few decades. Initially, tracking systems relied on basic Global Navigation Satellite Systems (GNSS) technology, which suffered from poor accuracy, high power consumption, and limited functionality in certain environments, such as indoors or in densely populated urban areas.
Moreover, GNSS systems are vulnerable to jamming and spoofing attacks, which can compromise their reliability for strategic applications.
Recent advancements in Radio Network-Based Positioning technology have addressed these issues by providing reliable positioning in situations where GNSS is less effective or when it is crucial to minimize power consumption in battery-powered devices.
Some of the key benefits of this technology include:
- The ability for companies to develop affordable asset management solutions that optimize battery life and ensure robust end-to-end security.
- A unified solution that supports asset management for both indoor and outdoor applications.
- The capability for customers to balance performance, location accuracy, and low power consumption is essential for battery-operated devices.
Basic Principles
Each cellular radio base station, known as a Cell, has a unique identifier called the Cell Identity (Cell ID). The same applies to WiFi Access Points (APs) and Bluetooth (BLE) beacons, which have a unique identifier called the MAC address that we utilize for indoor positioning.
Tracking devices like smartphones or IoT devices report the detected unique identifier (Cell IDs, WiFi APs, and/or BLE beacon IDs) to a server and their signal strengths.
If the device supports the Global Navigation Satellite System (GNSS), that information is also sent to the server for training.
Using these reports, a model for each transmitter can be calculated and stored in a database. Based on the measurement reports received for detected transmitters, this stored model can later be utilized to determine the positions of other devices lacking built-in GNSS capabilities.
Signal Strength and Quality Assessment
- Received Signal Strength Indicator (RSSI)
RSSI values indicate how strong the connection is between the device and each surrounding tower. Stronger signals suggest closer proximity. The Combain system interprets this RSSI data to adjust the estimated location. - Timing Advance (TA
The timing advance (TA) measured by the device and the network also provides an estimate of the distance between the cell tower and the device. This data helps to improve positioning accuracy, especially in rural areas where there is normally a line of sight (LOS) between the device and the cell tower. - 5G network timing measurements (TDOA & Round-Trip Time (RTT)
In 5G cellular networks, the network can make time measurements that are much more accurate than just using RSSI. Using the time difference of arrival (TDOA) for many base stations to a device and knowing the exact location of these base stations allows the location management function (LMF) to calculate an estimate of the device. These measurements are instrumental in indoor and urban settings where dense structures interfere with direct signal paths, allowing Combain’s algorithms to adjust for multipath interference (where signals bounce off buildings or walls).
Continuous self-improvement
Crowdsourcing
Service coverage and accuracy improve continually through the power of crowdsourced training data. As your applications or devices contribute data, the service rises above all others, achieving remarkable accuracy that sets a world-class standard.
-
Combain Crowdsourcing:
Gathers location data from a broad network of contributors to build and maintain a highly accurate and comprehensive location database. -
Real-Time Updates:
Ensures continuous, up-to-date information outperforming traditional static location data sources. -
Accuracy Through Diversity:
Combines data from various sources to create a more complete and accurate location picture. -
Empowerment and Collaboration:
Contributes to navigation, public safety, and smart city infrastructure innovations.
Combain crowdsourcing creates the foundation for the world’s most comprehensive and accurate location database. By leveraging the collective input of millions of users and devices, organizations can gather a wide range of real-time, geographically diverse data, outdoor and indoor. This approach enables continuous updates and refinements of location information, ensuring the database remains current and highly reliable.
The strength of crowdsourcing lies in its ability to aggregate data from countless sources, providing a multidimensional view of locations that static data collection methods cannot achieve. Contributors who share positioning data benefit from the enriched value of their location-aware applications, including mapping services, asset management, and emergency response systems. Furthermore, contributors may gain access to aggregated location insights, which can inform decision-making and foster innovations that benefit both individuals and industries.
The real value of crowdsourced positioning data comes from its accessibility and scalability. As more contributors share their information, the database expands exponentially in accuracy and coverage. This results in a dynamic, evolving location database consistently reflecting the most precise positioning information. Such a database is essential for applications requiring real-time location intelligence and supporting sectors such as security, telecommunications, transportation, urban planning, and public safety.
AI-Enhanced Positioning Technology
Radically improved accuracy from same infrastructure
Combain’s AI-enhanced positioning technology offers advanced location-based services through a sophisticated system that utilizes deep learning and neural networks. This system is designed to operate effectively indoors and outdoors by analyzing complex signal patterns from mobile network cell IDs, WiFi, Bluetooth, and LoRaWAN. This innovative method surpasses traditional trilateration techniques, resulting in a significant reduction in median error.
Unified Positioning Architecture
Combain’s AI architecture integrates extensive datasets, including cellular network data, WiFi access points, Bluetooth beacons, and LoRaWAN signals. Each signal source contributes additional detail, allowing the system to determine location accurately across various settings. This architecture also captures crowdsourced and user-reported data to enhance a global database of signal patterns, which forms the basis for training deep learning models.
The positioning models used by Combain are built on large-scale artificial neural networks (ANNs) that are optimized for unique signal signatures across different environments. By leveraging this global database, the system maintains adaptability, learning about regional variations while optimizing for different building materials and environmental factors. This flexibility enables Combain AI to operate efficiently in high-density urban areas, rural locations, and complex indoor spaces, where signal behavior can vary widely.
Read more about WiFi and Bluetooth indoor positioning here.
Data Collection and Processing
Data collection is a crucial aspect of Combain’s positioning technology, encompassing various signal types, including:
- Cellular Signals (2G, 3G, 4G, 5G)
These signals are collected globally to determine signal strength and cell ID patterns, essential for outdoor positioning. - WiFi Signals
The strength of each access point’s signal helps indicate a device’s proximity to known locations, which Combain uses for indoor positioning. - Bluetooth Beacons
Commonly utilized in buildings, these short-range signals provide enhanced accuracy for pinpointing locations in indoor spaces. - LoRaWAN Signals
Effective in areas with long-range, low-power IoT devices, these signals extend positioning capabilities in rural settings or large indoor areas.
The data is filtered and normalized to ensure accuracy and reliability. Signal variations are recorded at different times of day, across various geographic regions, and under different environmental conditions, allowing the system to accurately model real-world signal behavior.
Training the Neural Network Models
Combain’s AI models undergo rigorous training using supervised learning techniques. Each neural network is provided with labeled data consisting of signal attributes linked to known locations. This data helps the models map signal patterns to physical coordinates. The training process utilizes deep neural networks that capture real-world signal propagation’s non-linear and often unpredictable behaviors.
Combain’s team applies extensive pre-processing to normalize the data and eliminate any noise to ensure the network learns from the most relevant signal patterns. During the training phase, the models develop an understanding of spatial patterns, including how indoor signals are affected by building materials and how cellular signals vary across different terrains outdoors.
The models are adaptive, meaning they continually learn as new data becomes available, allowing them to maintain accuracy as environments and networks evolve.
This Android app is designed to survey and capture indoor training data. It has specific requirements for device capabilities, particularly regarding WiFi scanning and throttling. Please get in touch with us to confirm whether your equipment is compatible. We recommend using a Google Pixel 7 or later. Please note that the app is unavailable for iOS due to WiFi scanning limitations with iPhones.
SLAM for Indoor Navigation
Combain’s indoor positioning technology relies on Simultaneous Localization and Mapping (SLAM) algorithms, which map out spaces as users move through them, simultaneously locating themselves within these maps. By updating and refining internal maps in real-time, SLAM helps adjust for changes in signal sources (e.g., newly added Wi-Fi routers or changes in signal strength due to obstacles).
Operational Mechanism and Real-time Positioning
Once trained, Combain’s AI models are deployed through the Combain Location API, which provides real-time positioning services. The system uses signal inputs from the device to perform quick pattern-matching calculations within the model, accurately inferring the device’s location. For end-users, the positioning process is streamlined and requires only the available network data (e.g., in outdoor positioning using current and neighboring cell IDs or indoor the same for WiFi and BLE signals) to calculate the device’s location.
Combain’s positioning technology dynamically switches between models based on the available signals and the device’s environment. For instance, it may prioritize WiFi and Bluetooth signals in large indoor areas or utilize cellular and LoRaWAN networks in rural settings. This hybrid approach ensures that the system delivers consistent accuracy across various locations.
Read more about WiFi and Bluetooth indoor positioning here.
Technical Advantages of Combain’s AI Approach
- High Accuracy
Combain leverages advanced deep learning techniques to achieve impressive positioning accuracy, with error rates typically ranging from 1 to 5 meters indoors and varying more widely outdoors, depending on the region. - Low Power Consumption
The system is designed for IoT and mobile applications. It minimizes power usage by optimizing network calls, making it ideal for battery-operated devices. - Scalability
Combain’s architecture allows large-scale deployments, enabling organizations to implement the system across multiple buildings, cities, or regions without significant reconfiguration.
By integrating state-of-the-art deep learning methods with extensive data from diverse sources, Combain’s AI platform offers developers and engineers a powerful and versatile tool for creating location-based services that function seamlessly in indoor and outdoor environments.
Get started
Obtaining Your Key
To initiate the service, you must obtain a unique key from Combain. Depending on your license, this key will allow for either Cell ID or indoor positioning, or both. It’s important to keep this key confidential to prevent any misuse.
You can enable multiple keys for your account to differentiate between various services, models, and more.