Zebra Technologies Corporation · NASDAQ:ZBRA
Zebra Technologies — Zebra DevCon 2023
26 slides from Zebra Technologies's conference presentation, 2023
Slide 1: Leveraging Data Science to Improve Device Uptime
Leveraging Data Science to Improve Device Uptime Rob Reynolds Data Science Sr. Manager 1
Slide 2: Agenda
Leveraging Data Science to Improve Device Uptime Agenda Device uptime The steps The algorithm Applications of and battery Sense, analyze Remaining useful RUL performance and act life (RUL) of battery ZEBRA TECHNOLOGIES
Slide 3: Device Uptime and Battery
Leveraging Data Science to Improve Device Uptime Device Uptime and Battery Operational challenges that Battery challenges for the impact device uptime and This translates to: front-line workers availability • Can the battery last the shift • Front-line workers need to • Know your battery inventory –
Slide 4: The Steps
Leveraging Data Science to Improve Device Uptime The Steps Sense - Device data including battery related data coming from the battery pack. Consider any other relevant data from external sources Analyze - Go through the data and apply relevant algorithms to generate insights like Remaining Useful Li
Slide 5: Outcome - Smart Battery Report
Leveraging Data Science to Improve Device Uptime Outcome - Smart Battery Report A sample report to recommend batteries to be replaced based on RUL ZEBRA TECHNOLOGIES
Slide 6: Sense – Device and Battery Data Captured in the Cloud
Sense – Device and Battery Data Captured in the Cloud ZEBRA TECHNOLOGIES
Slide 7: Architecture Overview
Leveraging Data Science to Improve Device Uptime Architecture Overview ZEBRA TECHNOLOGIES
Slide 8: Data Available from Zebra Data Services (ZDS) agent
Leveraging Data Science to Improve Device Uptime Data Available from Zebra Data Services (ZDS) agent Along with battery related information, ZDS agent collects and transmits useful parameters like • Screen on time Zebra Data Services collects data on All these parameters impact battery • Screen brig
Slide 9: Defining Remaining Useful Life
Leveraging Data Science to Improve Device Uptime Defining Remaining Useful Life Either of the conditions When one of the could be met conditions is met Battery cannot receive Battery has undergone Battery may not be able Reliability of battery may or hold charge above a a threshold quantity of to pr
Slide 10: Defining Remaining Useful Life
Leveraging Data Science to Improve Device Uptime Defining Remaining Useful Life Factors that lead to battery degradation over time: • Battery Cells • Environment • Usage patterns Remaining Useful Life – The predicted number of days before the batteries key indicators fall below recommended levels Nu
Slide 11: The Input Data
Leveraging Data Science to Improve Device Uptime The Input Data Battery Data being considered • Data reported from the battery gauge – Battery Design Capacity – A measure of how much capacity a new battery of a particular type can deliver – Battery Present Capacity – a measure of how much charge the
Slide 12: Analyze – Battery RUL Algorithm
Analyze – Battery RUL Algorithm ZEBRA TECHNOLOGIES
Slide 13: The Process
Leveraging Data Science to Improve Device Uptime The Process Manufacturer Parameters for the battery for recommendations which RUL needs to be predicted Inputs VIQF Smart Cell Type Model Training Prediction Rated Capacity Battery Report We predict the remaining useful life in number of days Battery
Slide 14: Data Analysis and Data Quality
Leveraging Data Science to Improve Device Uptime Data Analysis and Data Quality Battery Present Capacity Battery Charge cycles should Battery should be uniquely should go down over a period be positive and monotonically identifiable of time. It cannot go up increasing defying laws of nature Battery
Slide 15: Training
Leveraging Data Science to Improve Device Uptime Training Reference - https://patentimages.storage.googleapis.com/65/52/6d/57ad11a747540e/US20220294027A1.pdf ZEBRA TECHNOLOGIES
Slide 16: The Model
Leveraging Data Science to Improve Device Uptime The Model Model gives values for degradation rates of health These degradation rates are then converted to number of days based on usage patterns The degradations are nonlinear due to varying environmental factors and battery cell characteristics ZEBR
Slide 17: Predictions
Leveraging Data Science to Improve Device Uptime Predictions Reference - https://patentimages.storage.googleapis.com/65/52/6d/57ad11a747540e/US20220294027A1.pdf ZEBRA TECHNOLOGIES
Slide 18: Sample Prediction Using the Model Output
Leveraging Data Science to Improve Device Uptime Sample Prediction Using the Model Output ZEBRA TECHNOLOGIES
Slide 19: Act – Proactive Battery Replacement Service
Act – Proactive Battery Replacement Service An application of RUL ZEBRA TECHNOLOGIES
Slide 20: PBR service
Leveraging Data Science to Improve Device Uptime PBR service For a Proactive Battery Replacement (PBR) customer, VIQ supporting the PBR service will facilitate battery orders to Siebel based on the current RUL reporting for customer who have batteries with an RUL < 30 days These batteries which are
Slide 21: Application of RUL – Planning Battery Inventory
Application of RUL – Planning Battery Inventory ZEBRA TECHNOLOGIES
Slide 22: Input Factors for Machine learning/simulations for battery
Leveraging Data Science to Improve Device Uptime Input Factors for Machine learning/simulations for battery Power Demand The battery Impact of battery Supply Profile - Profile - Demands present charge at degradation over Present capacity of Number of hours on the battery start of shift (the time on
Slide 23: Natural Degradation of Battery
Leveraging Data Science to Improve Device Uptime Natural Degradation of Battery Average Capacity Over Time ZEBRA TECHNOLOGIES
Slide 24: Demand and Supply Profiling Across Sites – A Sample
Leveraging Data Science to Improve Device Uptime Demand and Supply Profiling Across Sites – A Sample Demand (mAh) 3000 - 4000 < 1000 1000-2000 2000-3000 3000-4000 > 4000 32% 50% 15% 0.5% 0.01% Distribution of devices based on Demand This case has the ü Battery Swap Conditions Considered: possibility
Slide 25: Sample Recommendations to minimize disruptions
Sample Recommendations to minimize disruptions Actions to Minimize Business Disruptions Impact Q1 Q2 Q3 Q4 Potential Productivity Loss on busy days (in hours) Forecast if no action is taken and Devices are charged as per associates charging behavior 56 57 58 61 Forecast if no action is taken and Dev
Slide 26: Thank You
Thank You ZEBRA and the stylized Zebra head are trademarks of Zebra Technologies Corp., registered in many jurisdictions worldwide .All other trademarks are the property of their respective owners. ©2023 Zebra Technologies Corp. and/or its affiliates. All rights reserved.