Abstract
Magnetic Induction (MI) is an efficient wireless communication method to deploy operational internet of underground things (IoUT) for oil and gas reservoirs. The IoUT consists of underground things which are capable of sensing the underground environment and communicating with the surface. The MI-based IoUT enable many applications, such as monitoring of the oil rigs, optimized fracturing, and optimized extraction. Most of these applications are dependent on the location of the underground things and therefore require accurate localization techniques. The existing localization techniques for MI-based underground sensing networks are two-dimensional and do not characterize the achievable accuracy of the developed methods, which are both crucial and challenging tasks. Therefore, this paper proposes a novel three-dimensional (3D) localization technique based on Isometric scaling (Isomap) for future IoUT. Moreover, this paper also presents the closed-form expression of the Cramer Rao lower bound (CRLB) for the proposed technique, which takes into account the channel parameters of the underground magnetic-induction. The derived CRLB provides the suggestions for an MI-based underground localization system by associating the system parameters with the error trend. Numerical results demonstrate that localization accuracy is affected by different channel and networks parameters such as the number of underground things, ranging error variance, size of the coils, and the transmitting power. The root mean square error performance of the proposed technique shows that increase in the number of turns of the coils, transmitting power, and the number of anchors improves the performance. Results also show that the proposed technique is robust to the ranging error variance in the range of 10 to 30 %; however, a further increase in the ranging error variance does not allow to achieve acceptable accuracy. Also, the results show that the proposed technique achieves an average of 30 % better localization accuracy compare to the traditional methods.
The Challenge of Underground Localization
Monitoring oil rigs, optimizing hydraulic fracturing, and managing extraction require real-time data from underground sensors. However, the underground environment is "hostile" to conventional wireless technology. Standard Radio Frequency (RF) and Electromagnetic (EM) waves attenuate rapidly when passing through soil, rock, and water.
While Magnetic Induction (MI) has emerged as a viable alternative because it uses time-varying magnetic fields that penetrate soil more effectively, locating these sensors in 3D space remains a massive hurdle. Most existing localization techniques are limited to 2D and fail to account for the non-linear nature of signal propagation in complex, sparse underground networks. Furthermore, there was previously no theoretical benchmark to determine the maximum achievable accuracy for MI-based 3D localization.
Proposed Methodology: MI and Isomap
The authors propose a multi-step 3D localization strategy that leverages both hardware innovation and advanced manifold learning.
MI-Based Ranging
The system uses tri-directional coils to ensure omnidirectional communication coverage, minimizing the impact of coil misalignment. Distances between "things" are estimated based on the received signal strength, accounting for soil conductivity and "skin depth" effects.
Isomap (Isometric Scaling)
This is the core innovation. Unlike standard methods that assume simple straight-line (Euclidean) distances, Isomap treats the network as a "manifold." It uses Dijkstra’s algorithm to find the shortest paths between nodes to estimate missing pairwise distances, effectively capturing the non-linear structure of the underground network.
Procrustes Analysis and CRLB
Once a relative 3D map is created via Isomap, the researchers use Procrustes analysis—a linear transformation technique—to align these relative coordinates with absolute anchor positions. Additionally, the authors derived a closed-form expression for the Cramer Rao Lower Bound (CRLB). This provides a mathematical "gold standard" for the minimum possible error any localization algorithm can achieve given specific hardware and environmental parameters.
Key Results and Insights
The study relies on numerical simulations modeling a realistic 15 x 15 x 15 m³ cubic fracture at a depth of 1.8 km with 60 underground sensors.
- Superior Accuracy: The proposed Isomap technique achieved a Root Mean Square Error (RMSE) of 0.45 meters, vastly outperforming Local Linear Embedding (LLE) at 1.6m and Weighted-Multidimensional Scaling (WMDS) at 3.21m.
- 30% Improvement: On average, the technique provides a 30% boost in localization accuracy compared to traditional methods.
- Hardware Sensitivity: The derived CRLB shows that localization accuracy is highly sensitive to the number of turns in the induction coils and the transmit power. Accuracy improves significantly as the number of anchor nodes increases or as the coil size/turns are increased.
- Robustness to Noise: The algorithm remains highly stable even when ranging errors reach 30%, making it practical for the unpredictable conditions of a reservoir.
Conclusion
This research presents a robust solution for tracking sensors in deep underground oil and gas reservoirs. By combining Magnetic Induction communication with Isomap manifold learning, the framework provides high-accuracy 3D localization that significantly outperforms traditional methods. Future work will explore distributed localization architectures and the impact of heterogeneous soil properties on system performance.
Summary updated