Skip to main content

Proposal title SMART Wi-Fi : Management of the Wi-Fi Spectrum and Performance

Periodic Reporting for period 1 - Smart-WiFi (Proposal title SMART Wi-Fi : Management of the Wi-Fi Spectrum and Performance)

Reporting period: 2020-04-01 to 2021-03-31

The case for universal access to reliable, fast Internet is undeniable. With dramatic rises in demand, broadband networks have become bigger and more complex. Covid-19 and the shift to remote work and online education fast-forwarded the urgency for measures to improve quality and reliability of networks, in particular, in-home Wi-Fi networks. Lacking comprehensive tools to tackle interference and congestion problems on Wi-Fi networks, Internet service providers (ISPs) face mounting pressure on their bottom line, the risk of customer churn and related impact on their revenues.

This project aims to develop SMART-WIFI, an Artificial intelligence (AI) based spectrum broker solution to tackle problems associated with dense and uncoordinated Wi-Fi deployments, specifically addressing the most pressing challenge of radio frequency (RF) interference on Wi-Fi’s unlicensed spectrum. SMART-WIFI is a cutting-edge solution to interference problems, which is available to ISPs, enterprises and to end-users. Patented technologies underlying SMART-WIFI enable smart and dynamic channel allocation to respond instantly to interference from nearby wireless networks and devices, thereby substantially improving Wi-Fi performance and Quality of Experience (QoE). SMART-WIFI is extensively automated and scalable, as clients can instantly integrate the system without having to purchase hardware or install software updates on their in-home firmware. Clients can either integrate the Ambeent SDK into an application owned by their organization or download the Ambeent App through Google Play or Apple App Store. In this way, Ambeent’s technologies lessen the need for expensive Internet equipment by putting second hand, older access points (APs) into use and reliable networks can be extended to low-income households and areas.
Major achievements
• Ambeent signed contracts with three ISPs – TurkNet (>400.000 subscribers),Millenicom (>400.000) Vodafone (>1,000,000)–with a potential to reach 1.8 million Wi-Fi APs and 10 million Wi-Fi users.
• Filed 8 patents, which have led to the publication of peer-reviewed articles in top journals (authored by Ambeent employees).
• Published 2 White Papers, organized 2 webinars and appeared in 2 news feeds.
• Received recognition from leading industry organizations: Wireless Broadband Alliance, Hello Tomorrow, Telecom Council, Mobile World Congress, Red Herring.

Technical milestones
• Large scale demonstrations with Millenicom, TurkNet and Vodafone (WP 3). As a result, collected 113,039,210 measurements from 1,153,819 active users detecting 8,222,026 modems.
• TurkNet subscribers that use the TurkNet app have increased from about 20,000 to about 280,000 subscribers due to Ambeent’s optimization technology (Figure 1)
• Wi-Fi optimization was carried out for TurkNet resulting in average improvement of 52% in baseline performance by reducing neighborhood interference (Figure 2) ; link speed was improved by 280% on average (Figure 3); Wi-Fi speed was improved by 400% on average (Figure 4) (see Figure 6 for Wi-Fi speed definition). Also, see Figure 5 for data on customers that experienced the highest performance gains.
• Modem library expanded to 120 modems; app released on IOS and Android app stores (WP 2).
• Prediction models created for ISPs (WP 1, 2, 3).

WP1 ACTIVITIES: AI and Cloud Engine Platform Optimization
We fine-tuned our AI and cloud engine platforms with new algorithms to reduce processing delays between optimizations. We optimized channel distribution in crowded environments (4 APs average per router). The measurements we collected from 2 contracted ISPs allowed us to assess our application performance and construct prediction models. We developed real-time, hourly, and daily prediction models to create more accurate optimization profiles for users. Also, better predictive capabilities greatly helped ISPs to adjust their resources.
(See Table 1 and Figure 6,7,8 for detailed KPIs)

WP 2 ACTIVITIES: Build extensive modem libraries, enhance application
SMART-WIFI is currently interoperable with networks of 20 operators (10 in the EU) (Table 2). and supports 120 Wi-Fi makes and models (Figure 7,Table 3). Furthermore, our UIX interface has been updated based on a study of a sample of independent users to verify perceived needs. Our application is available in both IOS and Android app stores. Finally, we developed interfaces that can be integrated into ISPs’ call center portals with minimal coding requirements.

WP 3 ACTIVITIES: Large scale demonstration
We started pilot studies with 3 contracted ISPs and optimized the networks of more than 1,153,819 active users detecting 8,222,026 modems for two ISP clients (see monthly numbers in Table 4; Figure 9) .

WP 4 ACTIVITIES: Dissemination & Exploitation
We have sent offers to 5 ISP clients (covering a total of >15million APs), and we have a total of 5 PoCs in progress (Total >50 million APs) . We were in contact over email with about 100 different potential customers (Table 5).
Currently, there are various vendors that seek to address Wi-Fi problems resulting from channel interference: providers of smart routers (TP-link, Huawei, Zyxel, Netgear, Linksys), Wi-Fi Mesh Network Systems (Plume,Eero, Nest WiFi), SD-WAN companies (Cisco, Citrix, Riverbed, CloudGenix, VeloCloud, Viptela) and vendors providing tools for Network Performance Management and Diagnostics and Digital Experience Monitoring (Nyansa, Assia, 7Signal). Unlike SMART-WIFI, these solutions take a restricted approach in tackling RF interference causing significant delays in data transfers. Furthermore, all these solutions require new hardware and/or a software integration in the middleware gateway (i.e. APs) leading to complex and lengthy integration process.

SMART-WIFI’s unique selling points are:
User-centric: The patent pending technology allows for an API (installed in the user device) to collect detailed information including on neighboring APs and automatically assign the most efficient channel to a modem.
Collaborative, centralized spectrum management: All of the data collected by the user device is sent to the cloud to feed a central repository and machine-learning algorithms. Optimization is carried out in a collaborative way, after considering various parameters, including the real-time requirements of all Wi- Fi APs in a given cluster, thereby, optimizing channel allocation in that cluster.
User location and application aware: If for example, the user is using channel 6, and walks to the kitchen where the neighbor also uses channel 6, the algorithm will tell the device to move to another channel such as 11. SMART-WIFI is also application aware to take into account throughput requirements of each user or device in the home.
Zero hardware/software integration, scalable, vendor agnostic: SMART-WIFI is the industry’s first solution that is interoperable with all customer hardware brands/models.

SMART-WIFI’s potential impacts:
-Increasing connectivity of end users (including remote students and workers) by 99% through the implementation of SMART-WIFI by ISPs.
-Increasing Average Revenue Per User (ARPU) of ISPs by decreasing their costs.
-Elimination of customer dissatisfaction and resulting churn of ISPs.
-Contributing to circular economy performance by reducing the number of unnecessary AP replacements.
Figure 6 Ambeent new KPIs and client dashboard page
Figure 1 The number of TurkNet subscribers raising with SMART-WIFI optimization feature.
Table 2 Interoperability with operator networks; Table 3 Supported Wi-Fi make and models
Table 1 New KPI explanations
Table 4 Monthly measurements for TurkNet and Millenicom
Figure 8 Screenshot of client dashboard and end-user device analysis Samsung Xiaomi HUAWEI OPPO Gene
Figure 7 Screenshot of client dashboard and router analysis.
Figure 2 Fidelity improvement of selected TurkNet subscribers after SMART-WIFI optimization
Figure 5 Subscribers which experienced the most improvement with SMART-WIFI optimization
Figure 9 Measurements taken from users in different locations, covering more than 90% of the Turkey.
Figure 3 Link speed improvement of selected TurkNet subscribers after SMART-WIFI optimization
Table 5 Contacted customers
Figure 4 Wi-Fi speed improvement of selected TurkNet subscribers after SMART-WIFI optimization