Can We Enhance the Quality of Mobile Crowdsensing Data Without Ground Truth?

Jiajie Li,Bo Gu,Shimin Gong,Zhou Su,Mohsen Guizani

2024-05-29

Abstract:Mobile crowdsensing (MCS) has emerged as a prominent trend across various domains. However, ensuring the quality of the sensing data submitted by mobile users (MUs) remains a complex and challenging problem. To address this challenge, an advanced method is required to detect low-quality sensing data and identify malicious MUs that may disrupt the normal operations of an MCS system. Therefore, this article proposes a prediction- and reputation-based truth discovery (PRBTD) framework, which can separate low-quality data from high-quality data in sensing tasks. First, we apply a correlation-focused spatial-temporal transformer network to predict the ground truth of the input sensing data. Then, we extract the sensing errors of the data as features based on the prediction results to calculate the implications among the data. Finally, we design a reputation-based truth discovery (TD) module for identifying low-quality data with their implications. Given sensing data submitted by MUs, PRBTD can eliminate the data with heavy noise and identify malicious MUs with high accuracy. Extensive experimental results demonstrate that PRBTD outperforms the existing methods in terms of identification accuracy and data quality enhancement.

Machine Learning,Multiagent Systems

What problem does this paper attempt to address?

This paper attempts to solve the problems of data quality assessment and enhancement in Mobile Crowdsensing (MCS) systems. Specifically, the paper focuses on how to improve the sensing data quality of MCS systems and identify malicious users in the absence of ground truth.

### Background of the Paper's Problem

MCS systems collect a large amount of data through mobile devices (such as smart phones, cameras, etc.) and are widely used in various fields. However, due to sensor damage or the existence of malicious users, this sensing data may be unreliable. In addition, since this data is submitted by mobile users (MUs), the sensing platform lacks real - time labels, so it is difficult to directly assess data quality. This leads to the complexity and challenges of ensuring data quality in MCS systems.

### Proposed Method

To solve the above problems, the paper proposes a Prediction - and Reputation - based Truth Discovery (PRBTD) framework. The main purpose of this framework is to separate low - quality data from high - quality data and identify malicious users. The specific steps of the PRBTD framework are as follows:

1. **Prediction Module**:
- Use an improved Correlation - focused Spatial - Temporal Transformer Network (CFSTTN) to predict the real labels of the input sensing data.
- This network enhances its ability to extract spatio - temporal correlations through long - distance residual connections.

2. **Feature and Implication Calculation Module**:
- Extract the sensing error as a feature according to the prediction result.
- Calculate the implications between data at different locations and different time periods, and evaluate the data quality based on these features.

3. **Reputation - based Truth Discovery Module**:
- Use implications and users' reputations to infer real labels.
- Through alternately updating data quality and user reputations, finally converge to an accurate real - label estimate.

### Main Contributions

- **Combination of Prediction, Truth Discovery and Reputation Mechanisms**: For the first time, these three methods are combined to assess the quality of sensing data.
- **Handling Sparse Data**: By calculating the implications of data at different locations and different time periods, PRBTD can effectively assess data quality in sparse data scenarios.
- **Coping with Burst Values**: Overcome the limitations of prediction methods in dealing with burst values and improve the performance of MCS systems.

### Experimental Verification

The paper verifies the effectiveness of the PRBTD framework through a large number of experiments. The experimental results show that PRBTD is superior to existing methods in complex scenarios, can more accurately identify low - quality data and malicious users, and thus significantly improve the data quality of MCS systems.

In conclusion, this paper aims to propose a novel framework by combining prediction, truth discovery and reputation mechanisms to improve the quality of sensing data in MCS systems and effectively identify malicious users.