ML Methods and Anomaly Detection for dedicated 5G networks

5G (ENCQOR) Technology Development Challenge for SMEs

ML Methods and Anomaly Detection for dedicated 5G networks

Challenge Launch Date:

December 22nd, 2020

Challenge Deadline: 

January 31st,2020

Challenge Statement:

Industry 4.0 applications and services, such as autonomous driving, mining, logistics centers, ports, etc., require dedicated 5G networks with a sustained low latency and high bandwidth requirements.

These mission critical dedicated 5G networks will enable a wide range of time-critical services for consumers, enterprises and public institutions across various sectors. Such networks are intended for time-critical applications that demand data delivery with specific latency and reliability guarantees.

It is important that the reliability and latency constraints imposed in these mission critical dedicated 5G networks are constantly monitored and reinforced to maintain the quality of service (QoS), service level agreement (SLA) as infringements can cause safety threats and/or loss of income.

However, in practice, particularly in rugged outdoor environments, client devices experience signal fading and radio shadow due to new objects blocking the line of sight, and radio interference due to reuse of radio resources and from other client devices.

The objective is to improve the mission critical dedicated 5G network’s reliability and availability for demanding use-cases by applying modern ML algorithms and stream-analytics for QoS /SLA monitoring and prediction.

The goal of this challenge is to:

1. Design and develop ML methods to detect latency and reliability degradation and anomaly detection models for streaming data obtained from a dedicated 5G network (ENCQOR 5G)
2. Design and develop these models to comply with the latency requirements of the applications in a mission critical dedicated 5G network.
3. Design and propose a distributed execution framework for anomaly detection, correlation, prediction, and root cause analysis to provide inferences/results needed for the network automation and client applications.

Project Partner:

Ericsson Canada Inc.

Timeline:

9-12 months starting in 21Q1

Available Funding:

Up to 500 000$ CAD

Applicant Type:

SME with less 500 employees registered in Quebec

Location:

At SME location in Quebec with 5G testing in one of ENCQOR 5G Hubs in MTL or QC

Project Details:

ML Methods and Anomaly Detection for dedicated 5G networks.

Project Goals/Outcomes:

• Fully demonstratable ML models for anomaly detection, correlation, prediction, and root cause analysis in stream processing platform.
• Model execution time and complexity to meet the ultra-low latency requirements of mission critical dedicated 5G applications.

Applicant Capabilities:

• Experience Software development and data science in mission critical applications.
• Work with 5G systems and delivery experience of end-to-end applications.

How to participate:

• Each SME interested to participate to this challenge must have an Encqor iPaaS access (free) in order to understand the 5G technological stacks they will work with for the challenge. SMEs with no iPaaS access need to apply here and select ‘Ericsson call for project : ML Methods and Anomaly Detection for 5G networks’ in the first drop down menu
• Selected SMEs will be invited to discuss with the project owner to insure a speedy selection process