10/21/2021 | News release | Distributed by Public on 10/21/2021 20:36
Gamifying your app can potentially increase user engagement and incentivize your users to reach their goals as well as those of your organization. Now, as developers and others see the benefits, gamification is finding its way into IoT applications, including smart-city initiatives. Let's take a look at a few projects that illustrate this.
Many machine learning (ML) models rely on supervised learning, whereby the model is presented with data and the correct associated labels on which to train. Unfortunately, the process of labeling this data can be time consuming and difficult for humans, especially when it comes to associating sensor data that measures human activities. The result is often a lack of correctly labeled data or datasets that are too small to train ML models to the desired level of accuracy.
In their paper A Gamification Framework for Sensor Data Analytics, L'Heureux et Al. propose a socially engaging labeling system. Here the system automatically associates user actions (e.g., turning lights on and off) with sensor data, while incorporating gamification to motivate user's participation. The end goal was to see if their system could improve data labeling for supervised learning.
In a nutshell, the system merges gaming events (i.e., user actions) and sensor events (i.e., events detected in sensor data) and looks for associations between them for use in analytics. For example, when a user turns off a light, that event is associated with the changes in voltage and kilowatts detected by sensors. The user scans a QR code on the light switch to associate their action with sensor data so the system can automatically label the lighting event. A gamification framework is in place to encourage users to perform this scan. It includes a level hierarchy to unlock more actions and a leaderboard to show participation levels compared to other users. Of course, the real test would come when applying the system to a real-world use case to improve data labeling.
In the paper, the authors apply the framework at a power-metering company that sought to develop an ML-based sensor system to monitor room occupancy and lighting usage. The framework was used to incentivize users to record light switch events across different areas of the company's second floor.
The paper concluded that users positively changed their behavior and noted a near 17% increase in accuracy when training their ML models using the automated labeling system versus their best unsupervised approach. Anecdotally, they noted that gamification had a positive impact and almost eliminated wasted energy where the project was conducted.
This is a great example of how gamification can impact the development of IoT and AI systems. In this case, increased user participation helped associate user actions with sensor data, supporting the underlying goal of improving labeled training data and, ultimately, ML models with accuracy.
In A Biowaste Management Solution Stepped up an Octave, we looked at how the AxiBio Gaïabox rewards residents for ecologically responsible disposal of compostable items. The box measures the weight of the compostable waste, records it in the user's account, and uploads the data to the cloud. Since residents are charged for the amount of waste they dispose of, the weight recorded by the Gaïabox is translated into credits towards their overall garbage disposal fee.
The Gaïabox is another example of how incentivization in IoT can encourage and reward users for complying the eco-friendly behavior.
While it's easy to assume that gamification is something to introduce into an IoT system, researchers in Europe have designed a framework that works in the opposite direction. In this example, they utilized IoT data in serious games to encourage behavior changes.
In their paper InLife: Combining Real Life with Serious Games using IoT, Kosmides et Al. describe how their framework utilizes real-world data collected from an IoT system as input to a game. The framework's architecture comprised two main layers:
The authors describe how the framework was used as the basis for ICEBERG, an RPG for Android, where players are rewarded or penalized based on their actions. These actions are categorized as positive or negative, respectively, based on their environmental impact (e.g., turning off the lights when leaving the room, using excessive paper when printing, etc.). User actions are captured using various sensors placed in a real-world environment (e.g., an office). Then the data is communicated wirelessly to the game framework running either locally or in the cloud. As users engage in more eco-friendly actions, they're rewarded in the game with new accomplishments, missions, and resources to help them progress. Likewise, negative user actions count against these elements and hinder the user's progress. The game could be used for various purposes, such as encouraging energy-efficient behaviors in the workplace to ultimately reducing an organization's energy bill.
ICEBERG, and more generally the author's framework, illustrate how gameplay in a serious game can be linked to the real world. In this example, gameplay encourages users to engage in more eco-friendly behaviors in a fun yet serious way.
Gamification is now finding its way into new IoT use cases like smart city initiatives. We hope the examples above have inspired you to incorporate gamification into your next application. Perhaps one of the biggest takeaways is that incentivizing users is often more effective than penalizing them, which may be why gamification can be beneficial.
Qualcomm Technologies offers developers a number of platforms that can be used for building IoT applications and incorporating gamification systems. Below are a few notable examples:
If you need a primer on gamification, check out our blog, The Psychology and Benefits of Gamifying Apps. We discussed human factors that can make gamification successful, types of gamification methods, and key steps for implementing gamification.
Also, be sure to check out some of our recent blogs related to IoT and edge computing: