Data Analytics
UWB LOS & NLOS

31 March 2025



Project conducted as part of the Computing Science program at UofG and SIT





Description


The Context: Indoor localization (also known as an Indoor Positioning System or IPS) is a technology used to determine the precise location of people or objects inside a building. The project establishes a comprehensive data-driven framework designed to solve the problem of precise indoor localization using Ultra-Wideband (UWB) technology. This is because GPS signals are ineffective indoors due to attenuation. Hence, this project utilizes UWB Channel Impulse Response (CIR) measurements to accurately classify signals as Line-of-Sight (LOS) or Non-Line-of-Sight (NLOS) and predict measured ranges for improved positioning accuracy.



The Goal: This mini project aims to familiarize ourselves with the design, implementation and performance testing of the Line-of-Sight (LOS) & Non-Line- Sight(NLOS) Ultra-WideBand (UWB) wireless signal classification prediction.



Key Contributions

  • Developed & Trained (with hyper-tuning) the following models for the Classification of LOS/NLOS:
    • Random Forest Classifier
    • XG Boost Classifier
    • Logistic Regression Classifier
  • Developed & Trained (with hyper-tuning) the following models for the Range Estimation of signal:
    • Random Forest
    • XG Boost
    • Linear Regression
  • Data Mining
  • Data Visualizations



Tech Stack


  • Language: Python
  • Tools & Packages: Google Colab, matplotlib, sklearn, xgboost




Key Features & Functionality

Key functionalities include the derivation of three critical power metrics which enhance the characterization of signals under various environmental obstructions:

  • First Path Power Level
  • Received Power Level
  • Threshold Power Level


The system employs a sophisticated analytical pipeline featuring multi-stage feature selection (variance filtering, correlation analysis and mutual information ranking) and dimensionality reduction via Principal Component Analysis (PCA) to optimize data for machine learning.


To ensure high performance, the project implements various models, including Random Forest and XGBoost, which demonstrated superior ability in capturing the non-linear relationships inherent in complex indoor signal propagation.




Technical Challenges & Solutions



Technical Challenges & Solutions - Classifying LOS/NLOS


Issue

The primary technical hurdle is the presence of Non-Line-of-Sight (NLOS) conditions caused by physical obstructions like walls or doors. In these scenarios, the direct path between the UWB anchor and tag is blocked, leading to significant signal distortion and inaccurate range estimation, which ultimately degrades the precision of indoor localization systems. Traditional localization methods often struggle to distinguish between these distorted NLOS signals and clear LOS signals.



Solution

To mitigate the impact of NLOS conditions, our team developed a classification model leveraging advanced feature engineering. They systematically derived new signal metrics from existing UWB data, specifically the Threshold Power Level, which measures the difference between first path and received power levels to estimate signal attenuation. By combining these derived metrics with ensemble models like XGBoost and Random Forest, the solution achieved high classification accuracy (up to 91.67%), effectively identifying when a signal has been obstructed.



Newly Derived Feature Engineering Metrics
Random Forest Classifier Results
XGBoost Classifier Results



The Lesson

A lesson learnt from this challenge was that domain-specific feature engineering is essential when dealing with raw sensor data. That is to say, simply relying on predefined heuristics is often insufficient for complex environments. Instead, deriving metrics that directly represent physical phenomena, such as signal attenuation, provides the models with the necessary context to learn intricate patterns that simpler models might miss.




Technical Challenges & Solutions - Dealing with High Dimensionality


Issue

The raw dataset contained 1,016 samples of Channel Impulse Response (CIR) data for every measurement, presenting a high-dimensional data challenge. Many of these features were erratic, redundant, or non-informative, which increased the risk of model overfitting and created significant computational overhead, limiting the potential for real-time application.



Below is a small subset of the original dataset

first 3 rows of original dataset 
NLOS,RANGE,FP_IDX,FP_AMP1,FP_AMP2,FP_AMP3,STDEV_NOISE,CIR_PWR,MAX_NOISE,RXPACC,CH,FRAME_LEN,PREAM_LEN,BITRATE,PRFR,CIR0,CIR1,CIR2,CIR3,CIR4,CIR5,CIR6,CIR7,CIR8,CIR9,CIR10,CIR11,CIR12,CIR13,CIR14,CIR15,CIR16,CIR17,CIR18,CIR19,CIR20,CIR21,CIR22,CIR23,CIR24,CIR25,CIR26,CIR27,CIR28,CIR29,CIR30,CIR31,CIR32,CIR33,CIR34,CIR35,CIR36,CIR37,CIR38,CIR39,CIR40,CIR41,CIR42,CIR43,CIR44,CIR45,CIR46,CIR47,CIR48,CIR49,CIR50,CIR51,CIR52,CIR53,CIR54,CIR55,CIR56,CIR57,CIR58,CIR59,CIR60,CIR61,CIR62,CIR63,CIR64,CIR65,CIR66,CIR67,CIR68,CIR69,CIR70,CIR71,CIR72,CIR73,CIR74,CIR75,CIR76,CIR77,CIR78,CIR79,CIR80,CIR81,CIR82,CIR83,CIR84,CIR85,CIR86,CIR87,CIR88,CIR89,CIR90,CIR91,CIR92,CIR93,CIR94,CIR95,CIR96,CIR97,CIR98,CIR99,CIR100,CIR101,CIR102,CIR103,CIR104,CIR105,CIR106,CIR107,CIR108,CIR109,CIR110,CIR111,CIR112,CIR113,CIR114,CIR115,CIR116,CIR117,CIR118,CIR119,CIR120,CIR121,CIR122,CIR123,CIR124,CIR125,CIR126,CIR127,CIR128,CIR129,CIR130,CIR131,CIR132,CIR133,CIR134,CIR135,CIR136,CIR137,CIR138,CIR139,CIR140,CIR141,CIR142,CIR143,CIR144,CIR145,CIR146,CIR147,CIR148,CIR149,CIR150,CIR151,CIR152,CIR153,CIR154,CIR155,CIR156,CIR157,CIR158,CIR159,CIR160,CIR161,CIR162,CIR163,CIR164,CIR165,CIR166,CIR167,CIR168,CIR169,CIR170,CIR171,CIR172,CIR173,CIR174,CIR175,CIR176,CIR177,CIR178,CIR179,CIR180,CIR181,CIR182,CIR183,CIR184,CIR185,CIR186,CIR187,CIR188,CIR189,CIR190,CIR191,CIR192,CIR193,CIR194,CIR195,CIR196,CIR197,CIR198,CIR199,CIR200,CIR201,CIR202,CIR203,CIR204,CIR205,CIR206,CIR207,CIR208,CIR209,CIR210,CIR211,CIR212,CIR213,CIR214,CIR215,CIR216,CIR217,CIR218,CIR219,CIR220,CIR221,CIR222,CIR223,CIR224,CIR225,CIR226,CIR227,CIR228,CIR229,CIR230,CIR231,CIR232,CIR233,CIR234,CIR235,CIR236,CIR237,CIR238,CIR239,CIR240,CIR241,CIR242,CIR243,CIR244,CIR245,CIR246,CIR247,CIR248,CIR249,CIR250,CIR251,CIR252,CIR253,CIR254,CIR255,CIR256,CIR257,CIR258,CIR259,CIR260,CIR261,CIR262,CIR263,CIR264,CIR265,CIR266,CIR267,CIR268,CIR269,CIR270,CIR271,CIR272,CIR273,CIR274,CIR275,CIR276,CIR277,CIR278,CIR279,CIR280,CIR281,CIR282,CIR283,CIR284,CIR285,CIR286,CIR287,CIR288,CIR289,CIR290,CIR291,CIR292,CIR293,CIR294,CIR295,CIR296,CIR297,CIR298,CIR299,CIR300,CIR301,CIR302,CIR303,CIR304,CIR305,CIR306,CIR307,CIR308,CIR309,CIR310,CIR311,CIR312,CIR313,CIR314,CIR315,CIR316,CIR317,CIR318,CIR319,CIR320,CIR321,CIR322,CIR323,CIR324,CIR325,CIR326,CIR327,CIR328,CIR329,CIR330,CIR331,CIR332,CIR333,CIR334,CIR335,CIR336,CIR337,CIR338,CIR339,CIR340,CIR341,CIR342,CIR343,CIR344,CIR345,CIR346,CIR347,CIR348,CIR349,CIR350,CIR351,CIR352,CIR353,CIR354,CIR355,CIR356,CIR357,CIR358,CIR359,CIR360,CIR361,CIR362,CIR363,CIR364,CIR365,CIR366,CIR367,CIR368,CIR369,CIR370,CIR371,CIR372,CIR373,CIR374,CIR375,CIR376,CIR377,CIR378,CIR379,CIR380,CIR381,CIR382,CIR383,CIR384,CIR385,CIR386,CIR387,CIR388,CIR389,CIR390,CIR391,CIR392,CIR393,CIR394,CIR395,CIR396,CIR397,CIR398,CIR399,CIR400,CIR401,CIR402,CIR403,CIR404,CIR405,CIR406,CIR407,CIR408,CIR409,CIR410,CIR411,CIR412,CIR413,CIR414,CIR415,CIR416,CIR417,CIR418,CIR419,CIR420,CIR421,CIR422,CIR423,CIR424,CIR425,CIR426,CIR427,CIR428,CIR429,CIR430,CIR431,CIR432,CIR433,CIR434,CIR435,CIR436,CIR437,CIR438,CIR439,CIR440,CIR441,CIR442,CIR443,CIR444,CIR445,CIR446,CIR447,CIR448,CIR449,CIR450,CIR451,CIR452,CIR453,CIR454,CIR455,CIR456,CIR457,CIR458,CIR459,CIR460,CIR461,CIR462,CIR463,CIR464,CIR465,CIR466,CIR467,CIR468,CIR469,CIR470,CIR471,CIR472,CIR473,CIR474,CIR475,CIR476,CIR477,CIR478,CIR479,CIR480,CIR481,CIR482,CIR483,CIR484,CIR485,CIR486,CIR487,CIR488,CIR489,CIR490,CIR491,CIR492,CIR493,CIR494,CIR495,CIR496,CIR497,CIR498,CIR499,CIR500,CIR501,CIR502,CIR503,CIR504,CIR505,CIR506,CIR507,CIR508,CIR509,CIR510,CIR511,CIR512,CIR513,CIR514,CIR515,CIR516,CIR517,CIR518,CIR519,CIR520,CIR521,CIR522,CIR523,CIR524,CIR525,CIR526,CIR527,CIR528,CIR529,CIR530,CIR531,CIR532,CIR533,CIR534,CIR535,CIR536,CIR537,CIR538,CIR539,CIR540,CIR541,CIR542,CIR543,CIR544,CIR545,CIR546,CIR547,CIR548,CIR549,CIR550,CIR551,CIR552,CIR553,CIR554,CIR555,CIR556,CIR557,CIR558,CIR559,CIR560,CIR561,CIR562,CIR563,CIR564,CIR565,CIR566,CIR567,CIR568,CIR569,CIR570,CIR571,CIR572,CIR573,CIR574,CIR575,CIR576,CIR577,CIR578,CIR579,CIR580,CIR581,CIR582,CIR583,CIR584,CIR585,CIR586,CIR587,CIR588,CIR589,CIR590,CIR591,CIR592,CIR593,CIR594,CIR595,CIR596,CIR597,CIR598,CIR599,CIR600,CIR601,CIR602,CIR603,CIR604,CIR605,CIR606,CIR607,CIR608,CIR609,CIR610,CIR611,CIR612,CIR613,CIR614,CIR615,CIR616,CIR617,CIR618,CIR619,CIR620,CIR621,CIR622,CIR623,CIR624,CIR625,CIR626,CIR627,CIR628,CIR629,CIR630,CIR631,CIR632,CIR633,CIR634,CIR635,CIR636,CIR637,CIR638,CIR639,CIR640,CIR641,CIR642,CIR643,CIR644,CIR645,CIR646,CIR647,CIR648,CIR649,CIR650,CIR651,CIR652,CIR653,CIR654,CIR655,CIR656,CIR657,CIR658,CIR659,CIR660,CIR661,CIR662,CIR663,CIR664,CIR665,CIR666,CIR667,CIR668,CIR669,CIR670,CIR671,CIR672,CIR673,CIR674,CIR675,CIR676,CIR677,CIR678,CIR679,CIR680,CIR681,CIR682,CIR683,CIR684,CIR685,CIR686,CIR687,CIR688,CIR689,CIR690,CIR691,CIR692,CIR693,CIR694,CIR695,CIR696,CIR697,CIR698,CIR699,CIR700,CIR701,CIR702,CIR703,CIR704,CIR705,CIR706,CIR707,CIR708,CIR709,CIR710,CIR711,CIR712,CIR713,CIR714,CIR715,CIR716,CIR717,CIR718,CIR719,CIR720,CIR721,CIR722,CIR723,CIR724,CIR725,CIR726,CIR727,CIR728,CIR729,CIR730,CIR731,CIR732,CIR733,CIR734,CIR735,CIR736,CIR737,CIR738,CIR739,CIR740,CIR741,CIR742,CIR743,CIR744,CIR745,CIR746,CIR747,CIR748,CIR749,CIR750,CIR751,CIR752,CIR753,CIR754,CIR755,CIR756,CIR757,CIR758,CIR759,CIR760,CIR761,CIR762,CIR763,CIR764,CIR765,CIR766,CIR767,CIR768,CIR769,CIR770,CIR771,CIR772,CIR773,CIR774,CIR775,CIR776,CIR777,CIR778,CIR779,CIR780,CIR781,CIR782,CIR783,CIR784,CIR785,CIR786,CIR787,CIR788,CIR789,CIR790,CIR791,CIR792,CIR793,CIR794,CIR795,CIR796,CIR797,CIR798,CIR799,CIR800,CIR801,CIR802,CIR803,CIR804,CIR805,CIR806,CIR807,CIR808,CIR809,CIR810,CIR811,CIR812,CIR813,CIR814,CIR815,CIR816,CIR817,CIR818,CIR819,CIR820,CIR821,CIR822,CIR823,CIR824,CIR825,CIR826,CIR827,CIR828,CIR829,CIR830,CIR831,CIR832,CIR833,CIR834,CIR835,CIR836,CIR837,CIR838,CIR839,CIR840,CIR841,CIR842,CIR843,CIR844,CIR845,CIR846,CIR847,CIR848,CIR849,CIR850,CIR851,CIR852,CIR853,CIR854,CIR855,CIR856,CIR857,CIR858,CIR859,CIR860,CIR861,CIR862,CIR863,CIR864,CIR865,CIR866,CIR867,CIR868,CIR869,CIR870,CIR871,CIR872,CIR873,CIR874,CIR875,CIR876,CIR877,CIR878,CIR879,CIR880,CIR881,CIR882,CIR883,CIR884,CIR885,CIR886,CIR887,CIR888,CIR889,CIR890,CIR891,CIR892,CIR893,CIR894,CIR895,CIR896,CIR897,CIR898,CIR899,CIR900,CIR901,CIR902,CIR903,CIR904,CIR905,CIR906,CIR907,CIR908,CIR909,CIR910,CIR911,CIR912,CIR913,CIR914,CIR915,CIR916,CIR917,CIR918,CIR919,CIR920,CIR921,CIR922,CIR923,CIR924,CIR925,CIR926,CIR927,CIR928,CIR929,CIR930,CIR931,CIR932,CIR933,CIR934,CIR935,CIR936,CIR937,CIR938,CIR939,CIR940,CIR941,CIR942,CIR943,CIR944,CIR945,CIR946,CIR947,CIR948,CIR949,CIR950,CIR951,CIR952,CIR953,CIR954,CIR955,CIR956,CIR957,CIR958,CIR959,CIR960,CIR961,CIR962,CIR963,CIR964,CIR965,CIR966,CIR967,CIR968,CIR969,CIR970,CIR971,CIR972,CIR973,CIR974,CIR975,CIR976,CIR977,CIR978,CIR979,CIR980,CIR981,CIR982,CIR983,CIR984,CIR985,CIR986,CIR987,CIR988,CIR989,CIR990,CIR991,CIR992,CIR993,CIR994,CIR995,CIR996,CIR997,CIR998,CIR999,CIR1000,CIR1001,CIR1002,CIR1003,CIR1004,CIR1005,CIR1006,CIR1007,CIR1008,CIR1009,CIR1010,CIR1011,CIR1012,CIR1013,CIR1014,CIR1015
0.00,3.90,745.00,18712.00,10250.00,11576.00,64.00,11855.00,967.00,611.00,2.00,39.00,1024.00,110.00,64.00,129.00,312.00,331.00,141.00,160.00,488.00,195.00,144.00,119.00,193.00,158.00,192.00,120.00,35.00,295.00,111.00,292.00,170.00,84.00,107.00,293.00,200.00,333.00,450.00,241.00,162.00,97.00,449.00,146.00,66.00,165.00,455.00,190.00,142.00,218.00,280.00,219.00,238.00,265.00,207.00,123.00,164.00,195.00,322.00,222.00,200.00,80.00,362.00,218.00,249.00,167.00,207.00,315.00,130.00,114.00,248.00,196.00,151.00,284.00,513.00,213.00,260.00,77.00,405.00,191.00,307.00,130.00,262.00,277.00,523.00,182.00,216.00,247.00,227.00,342.00,407.00,230.00,51.00,247.00,401.00,193.00,211.00,253.00,244.00,117.00,226.00,214.00,123.00,258.00,335.00,173.00,183.00,191.00,186.00,409.00,252.00,238.00,169.00,104.00,224.00,119.00,204.00,251.00,236.00,185.00,94.00,166.00,252.00,86.00,227.00,183.00,454.00,214.00,165.00,135.00,78.00,172.00,192.00,257.00,291.00,532.00,146.00,183.00,153.00,346.00,146.00,149.00,141.00,177.00,48.00,138.00,270.00,151.00,307.00,426.00,170.00,223.00,100.00,279.00,126.00,346.00,261.00,186.00,402.00,475.00,249.00,35.00,201.00,182.00,242.00,296.00,150.00,293.00,209.00,212.00,277.00,145.00,173.00,83.00,442.00,480.00,161.00,230.00,214.00,231.00,540.00,87.00,72.00,261.00,230.00,192.00,188.00,75.00,256.00,274.00,447.00,185.00,215.00,63.00,514.00,205.00,450.00,306.00,308.00,193.00,156.00,192.00,230.00,110.00,252.00,263.00,222.00,172.00,171.00,238.00,202.00,93.00,133.00,397.00,219.00,193.00,243.00,279.00,252.00,175.00,195.00,504.00,219.00,315.00,214.00,317.00,390.00,154.00,86.00,348.00,214.00,108.00,173.00,126.00,139.00,13.00,182.00,111.00,418.00,131.00,202.00,236.00,383.00,82.00,94.00,86.00,301.00,100.00,196.00,32.00,210.00,231.00,286.00,214.00,321.00,220.00,305.00,197.00,232.00,298.00,169.00,284.00,252.00,254.00,111.00,178.00,233.00,258.00,101.00,272.00,38.00,248.00,81.00,222.00,216.00,69.00,166.00,181.00,217.00,138.00,285.00,244.00,500.00,288.00,276.00,207.00,496.00,177.00,315.00,136.00,296.00,360.00,125.00,156.00,309.00,260.00,242.00,331.00,492.00,61.00,136.00,222.00,70.00,144.00,158.00,272.00,42.00,278.00,164.00,146.00,59.00,264.00,272.00,220.00,229.00,211.00,53.00,257.00,136.00,239.00,242.00,244.00,269.00,117.00,230.00,279.00,247.00,314.00,300.00,193.00,270.00,140.00,245.00,227.00,162.00,287.00,41.00,232.00,229.00,187.00,153.00,150.00,296.00,175.00,224.00,137.00,122.00,219.00,182.00,148.00,223.00,216.00,301.00,240.00,166.00,152.00,154.00,273.00,253.00,344.00,171.00,251.00,97.00,222.00,179.00,215.00,289.00,203.00,302.00,170.00,190.00,138.00,84.00,273.00,135.00,253.00,215.00,116.00,271.00,331.00,164.00,293.00,104.00,89.00,359.00,366.00,229.00,270.00,454.00,302.00,78.00,92.00,390.00,137.00,284.00,93.00,228.00,252.00,211.00,265.00,233.00,150.00,240.00,336.00,241.00,464.00,436.00,240.00,248.00,245.00,191.00,224.00,119.00,251.00,126.00,140.00,271.00,266.00,246.00,86.00,291.00,61.00,192.00,280.00,138.00,264.00,253.00,76.00,275.00,365.00,137.00,264.00,198.00,364.00,183.00,110.00,88.00,407.00,276.00,272.00,160.00,212.00,159.00,35.00,97.00,211.00,294.00,230.00,253.00,195.00,185.00,221.00,222.00,112.00,445.00,425.00,179.00,183.00,171.00,83.00,123.00,191.00,220.00,175.00,114.00,120.00,308.00,332.00,196.00,184.00,412.00,159.00,156.00,245.00,193.00,255.00,111.00,245.00,82.00,102.00,271.00,112.00,163.00,43.00,184.00,193.00,103.00,232.00,127.00,136.00,246.00,189.00,263.00,296.00,284.00,166.00,182.00,254.00,304.00,174.00,119.00,152.00,186.00,236.00,231.00,210.00,115.00,192.00,308.00,272.00,305.00,93.00,241.00,198.00,44.00,214.00,101.00,71.00,419.00,414.00,182.00,109.00,170.00,243.00,137.00,153.00,168.00,169.00,203.00,541.00,677.00,232.00,50.00,235.00,238.00,208.00,388.00,143.00,394.00,282.00,179.00,221.00,118.00,210.00,96.00,165.00,151.00,281.00,123.00,266.00,260.00,133.00,292.00,236.00,328.00,238.00,187.00,457.00,185.00,75.00,205.00,159.00,225.00,254.00,168.00,168.00,244.00,310.00,72.00,404.00,146.00,241.00,237.00,177.00,256.00,96.00,206.00,229.00,198.00,246.00,540.00,508.00,215.00,260.00,296.00,55.00,173.00,243.00,66.00,422.00,106.00,195.00,116.00,249.00,90.00,159.00,73.00,313.00,446.00,209.00,97.00,204.00,229.00,79.00,161.00,260.00,478.00,98.00,188.00,174.00,82.00,240.00,101.00,167.00,162.00,449.00,230.00,313.00,349.00,60.00,351.00,291.00,522.00,341.00,45.00,176.00,93.00,200.00,392.00,299.00,524.00,146.00,126.00,505.00,285.00,208.00,341.00,184.00,354.00,173.00,69.00,126.00,194.00,196.00,421.00,246.00,171.00,134.00,262.00,252.00,175.00,266.00,123.00,244.00,168.00,153.00,118.00,36.00,191.00,228.00,166.00,301.00,198.00,250.00,164.00,256.00,266.00,132.00,219.00,109.00,413.00,26.00,167.00,250.00,187.00,79.00,173.00,322.00,181.00,67.00,179.00,227.00,126.00,319.00,404.00,285.00,281.00,326.00,254.00,103.00,161.00,32.00,434.00,282.00,190.00,87.00,123.00,161.00,107.00,208.00,141.00,131.00,520.00,247.00,183.00,207.00,221.00,256.00,262.00,476.00,85.00,306.00,178.00,310.00,201.00,174.00,214.00,312.00,204.00,292.00,234.00,190.00,330.00,346.00,185.00,206.00,451.00,128.00,157.00,200.00,188.00,230.00,252.00,377.00,301.00,155.00,203.00,227.00,174.00,292.00,91.00,151.00,271.00,246.00,59.00,225.00,86.00,232.00,251.00,195.00,314.00,98.00,214.00,149.00,484.00,266.00,133.00,262.00,814.00,5242.00,11573.00,6102.00,12961.00,23077.00,9621.00,8303.00,7763.00,2337.00,6569.00,8875.00,6793.00,7093.00,8150.00,3271.00,5916.00,2278.00,6750.00,6027.00,351.00,4700.00,4515.00,1763.00,3388.00,1364.00,2288.00,1830.00,1307.00,461.00,1152.00,1289.00,798.00,400.00,640.00,1203.00,1252.00,714.00,776.00,495.00,531.00,528.00,480.00,644.00,268.00,473.00,421.00,553.00,1104.00,678.00,253.00,343.00,50.00,563.00,496.00,263.00,161.00,163.00,217.00,274.00,157.00,111.00,233.00,229.00,125.00,76.00,74.00,239.00,240.00,233.00,219.00,232.00,108.00,166.00,448.00,402.00,191.00,133.00,77.00,421.00,236.00,161.00,116.00,82.00,40.00,255.00,147.00,58.00,116.00,215.00,276.00,325.00,250.00,124.00,296.00,225.00,412.00,198.00,237.00,299.00,164.00,296.00,230.00,487.00,249.00,56.00,241.00,238.00,179.00,63.00,251.00,486.00,494.00,290.00,297.00,232.00,315.00,234.00,74.00,215.00,217.00,269.00,259.00,61.00,230.00,250.00,119.00,263.00,251.00,215.00,184.00,256.00,225.00,181.00,345.00,325.00,142.00,253.00,197.00,252.00,119.00,101.00,117.00,361.00,265.00,98.00,237.00,202.00,310.00,250.00,297.00,169.00,475.00,246.00,271.00,197.00,222.00,311.00,220.00,175.00,99.00,147.00,110.00,179.00,62.00,293.00,410.00,471.00,295.00,94.00,47.00,142.00,226.00,221.00,208.00,131.00,123.00,198.00,306.00,201.00,236.00,158.00,168.00,346.00,405.00,268.00,226.00,161.00,286.00,238.00,163.00,248.00,211.00,236.00,145.00,122.00,417.00,218.00,358.00,501.00,115.00,143.00,149.00,245.00,250.00,292.00,213.00,117.00,334.00,180.00,77.00,273.00,325.00,265.00,130.00,256.00,199.00,200.00,183.00,278.00,432.00,211.00,263.00,105.00,402.00,320.00,133.00,170.00,207.00,280.00,399.00,457.00,162.00,329.00,162.00,325.00,121.00,110.00,291.00,301.00,259.00,251.00,156.00,200.00,287.00,177.00,402.00,216.00,242.00,249.00,420.00,306.00,86.00,278.00,487.00,445.00,169.00,243.00,425.00,235.00,164.00,297.00,114.00,279.00,458.00,183.00,158.00,198.00,87.00,296.00,505.00,307.00,0.00
0.00,0.66,749.00,11239.00,6313.00,4712.00,64.00,18968.00,1133.00,447.00,2.00,27.00,1024.00,110.00,64.00,117.00,163.00,199.00,136.00,142.00,182.00,358.00,123.00,280.00,311.00,236.00,166.00,240.00,286.00,117.00,242.00,250.00,318.00,204.00,259.00,82.00,207.00,145.00,255.00,201.00,179.00,255.00,50.00,201.00,184.00,178.00,279.00,99.00,289.00,190.00,93.00,289.00,329.00,164.00,132.00,145.00,195.00,131.00,76.00,85.00,187.00,280.00,232.00,292.00,174.00,161.00,141.00,201.00,218.00,168.00,151.00,131.00,229.00,326.00,219.00,52.00,180.00,91.00,297.00,347.00,236.00,264.00,117.00,88.00,232.00,435.00,511.00,221.00,180.00,291.00,253.00,147.00,272.00,50.00,331.00,132.00,291.00,345.00,241.00,286.00,170.00,350.00,184.00,248.00,295.00,266.00,280.00,159.00,144.00,24.00,155.00,140.00,132.00,292.00,195.00,271.00,33.00,226.00,282.00,234.00,166.00,109.00,36.00,224.00,100.00,389.00,398.00,121.00,145.00,125.00,183.00,254.00,138.00,42.00,70.00,138.00,95.00,220.00,103.00,146.00,159.00,177.00,184.00,273.00,279.00,326.00,373.00,222.00,214.00,241.00,117.00,314.00,104.00,158.00,70.00,336.00,56.00,171.00,64.00,267.00,211.00,275.00,105.00,60.00,227.00,201.00,233.00,249.00,122.00,185.00,98.00,268.00,157.00,531.00,214.00,246.00,181.00,166.00,162.00,103.00,155.00,249.00,425.00,536.00,148.00,220.00,206.00,153.00,157.00,217.00,129.00,219.00,100.00,213.00,71.00,157.00,177.00,238.00,228.00,306.00,282.00,164.00,281.00,62.00,257.00,311.00,235.00,56.00,184.00,224.00,125.00,253.00,240.00,241.00,283.00,200.00,182.00,233.00,211.00,296.00,313.00,79.00,62.00,220.00,146.00,291.00,219.00,184.00,116.00,188.00,134.00,302.00,308.00,107.00,144.00,58.00,315.00,25.00,249.00,153.00,88.00,169.00,127.00,220.00,153.00,175.00,262.00,108.00,236.00,79.00,335.00,61.00,113.00,245.00,257.00,269.00,145.00,181.00,156.00,251.00,46.00,230.00,115.00,269.00,325.00,147.00,102.00,188.00,215.00,192.00,428.00,262.00,210.00,193.00,138.00,340.00,246.00,144.00,219.00,248.00,215.00,182.00,204.00,156.00,123.00,341.00,144.00,306.00,155.00,138.00,75.00,185.00,166.00,79.00,84.00,327.00,193.00,105.00,507.00,261.00,115.00,141.00,417.00,447.00,262.00,84.00,205.00,306.00,165.00,276.00,248.00,22.00,222.00,260.00,100.00,161.00,18.00,118.00,262.00,173.00,167.00,176.00,80.00,200.00,15.00,82.00,261.00,185.00,229.00,246.00,314.00,131.00,132.00,307.00,31.00,265.00,136.00,131.00,114.00,94.00,341.00,263.00,492.00,316.00,153.00,73.00,139.00,282.00,247.00,93.00,131.00,65.00,259.00,76.00,118.00,211.00,231.00,288.00,200.00,249.00,272.00,333.00,330.00,258.00,123.00,299.00,203.00,121.00,228.00,321.00,360.00,81.00,251.00,304.00,98.00,212.00,88.00,252.00,34.00,257.00,250.00,276.00,504.00,378.00,269.00,288.00,210.00,84.00,59.00,288.00,221.00,261.00,260.00,331.00,237.00,285.00,251.00,310.00,242.00,258.00,246.00,126.00,217.00,209.00,179.00,142.00,183.00,342.00,153.00,127.00,182.00,184.00,43.00,526.00,361.00,113.00,114.00,126.00,164.00,144.00,174.00,318.00,144.00,174.00,242.00,221.00,236.00,252.00,69.00,187.00,181.00,245.00,266.00,150.00,187.00,65.00,248.00,241.00,258.00,123.00,79.00,164.00,249.00,115.00,214.00,34.00,290.00,136.00,99.00,184.00,233.00,159.00,50.00,101.00,224.00,229.00,168.00,246.00,87.00,244.00,86.00,143.00,200.00,118.00,69.00,250.00,96.00,198.00,153.00,196.00,225.00,330.00,143.00,151.00,130.00,126.00,169.00,276.00,182.00,64.00,242.00,260.00,210.00,114.00,196.00,302.00,216.00,166.00,67.00,254.00,177.00,37.00,249.00,189.00,581.00,19.00,199.00,213.00,203.00,97.00,253.00,77.00,108.00,311.00,218.00,323.00,52.00,236.00,259.00,189.00,228.00,213.00,94.00,23.00,61.00,61.00,191.00,148.00,86.00,204.00,90.00,192.00,254.00,148.00,172.00,244.00,312.00,302.00,130.00,116.00,162.00,277.00,236.00,209.00,487.00,425.00,223.00,223.00,137.00,311.00,71.00,171.00,336.00,303.00,115.00,156.00,142.00,115.00,272.00,246.00,79.00,190.00,139.00,321.00,220.00,303.00,217.00,214.00,290.00,272.00,320.00,303.00,140.00,75.00,103.00,65.00,72.00,195.00,271.00,249.00,273.00,209.00,257.00,268.00,55.00,174.00,147.00,219.00,193.00,214.00,172.00,146.00,80.00,304.00,269.00,37.00,159.00,184.00,173.00,130.00,245.00,262.00,236.00,119.00,103.00,128.00,102.00,110.00,228.00,92.00,132.00,112.00,150.00,100.00,229.00,245.00,49.00,257.00,248.00,112.00,86.00,153.00,116.00,26.00,252.00,163.00,269.00,237.00,335.00,327.00,161.00,210.00,308.00,234.00,193.00,54.00,86.00,148.00,183.00,480.00,266.00,263.00,286.00,237.00,123.00,50.00,69.00,157.00,346.00,148.00,211.00,235.00,167.00,73.00,116.00,145.00,283.00,245.00,113.00,244.00,112.00,234.00,253.00,78.00,429.00,284.00,153.00,253.00,125.00,160.00,230.00,222.00,110.00,218.00,61.00,227.00,192.00,82.00,327.00,154.00,141.00,295.00,109.00,326.00,290.00,237.00,208.00,220.00,213.00,214.00,262.00,259.00,309.00,86.00,92.00,168.00,533.00,285.00,96.00,129.00,275.00,208.00,223.00,213.00,91.00,107.00,35.00,445.00,113.00,66.00,324.00,185.00,53.00,156.00,247.00,178.00,188.00,106.00,151.00,133.00,240.00,230.00,119.00,264.00,270.00,195.00,96.00,472.00,184.00,235.00,207.00,104.00,124.00,230.00,222.00,117.00,261.00,117.00,241.00,186.00,115.00,101.00,290.00,295.00,211.00,184.00,62.00,240.00,166.00,309.00,169.00,247.00,290.00,399.00,268.00,381.00,190.00,193.00,126.00,278.00,349.00,167.00,227.00,288.00,251.00,10.00,51.00,92.00,179.00,506.00,123.00,194.00,271.00,384.00,1840.00,5244.00,10885.00,17365.00,17764.00,12565.00,16559.00,17830.00,18059.00,14755.00,12559.00,13604.00,4861.00,6611.00,11470.00,6046.00,2636.00,3473.00,4099.00,7353.00,4080.00,910.00,997.00,1262.00,1124.00,4716.00,6172.00,3300.00,1037.00,1291.00,541.00,1510.00,2275.00,962.00,1916.00,2752.00,1195.00,1051.00,385.00,700.00,708.00,176.00,495.00,577.00,515.00,772.00,504.00,618.00,473.00,702.00,418.00,760.00,1001.00,495.00,1113.00,975.00,416.00,264.00,996.00,1003.00,477.00,472.00,339.00,439.00,472.00,300.00,666.00,501.00,366.00,478.00,332.00,251.00,226.00,515.00,573.00,113.00,289.00,104.00,211.00,137.00,77.00,173.00,173.00,125.00,126.00,105.00,202.00,138.00,312.00,243.00,215.00,215.00,185.00,236.00,350.00,186.00,150.00,135.00,198.00,310.00,267.00,294.00,218.00,312.00,53.00,129.00,273.00,249.00,235.00,267.00,348.00,192.00,106.00,284.00,222.00,97.00,155.00,129.00,307.00,316.00,116.00,244.00,153.00,309.00,321.00,234.00,232.00,40.00,233.00,193.00,182.00,200.00,258.00,139.00,64.00,148.00,232.00,389.00,423.00,342.00,129.00,38.00,25.00,278.00,89.00,402.00,470.00,272.00,265.00,265.00,249.00,269.00,161.00,233.00,239.00,257.00,265.00,266.00,55.00,46.00,128.00,145.00,187.00,127.00,163.00,94.00,328.00,166.00,46.00,218.00,236.00,299.00,219.00,252.00,143.00,265.00,309.00,247.00,299.00,67.00,56.00,97.00,119.00,307.00,169.00,201.00,114.00,193.00,128.00,234.00,225.00,258.00,30.00,187.00,393.00,339.00,221.00,161.00,296.00,59.00,243.00,465.00,244.00,343.00,284.00,110.00,228.00,289.00,202.00,285.00,151.00,254.00,200.00,122.00,81.00,223.00,75.00,174.00,263.00,249.00,148.00,187.00,229.00,172.00,107.00,65.00,338.00,147.00,51.00,283.00,255.00,322.00,311.00,159.00,330.00,194.00,57.00,148.00,123.00,149.00,62.00,60.00,379.00,196.00,140.00,90.00,177.00,257.00,277.00,87.00,473.00,277.00,107.00,104.00,161.00,168.00,196.00,144.00,334.00,290.00,228.00,187.00,213.00,202.00,89.00,103.00,0.00
1.00,7.86,746.00,4355.00,5240.00,3478.00,60.00,14699.00,894.00,723.00,2.00,27.00,1024.00,110.00,64.00,433.00,240.00,233.00,81.00,143.00,506.00,119.00,272.00,248.00,233.00,233.00,95.00,255.00,189.00,142.00,274.00,131.00,253.00,115.00,102.00,50.00,417.00,239.00,40.00,218.00,161.00,235.00,165.00,153.00,262.00,146.00,240.00,258.00,101.00,228.00,204.00,147.00,130.00,201.00,285.00,209.00,168.00,115.00,246.00,274.00,211.00,51.00,234.00,295.00,215.00,270.00,77.00,160.00,304.00,241.00,180.00,284.00,307.00,128.00,405.00,252.00,464.00,478.00,304.00,134.00,117.00,33.00,248.00,91.00,235.00,243.00,141.00,70.00,229.00,244.00,389.00,238.00,253.00,231.00,263.00,262.00,287.00,179.00,120.00,159.00,183.00,46.00,70.00,160.00,399.00,225.00,194.00,160.00,193.00,261.00,201.00,242.00,160.00,182.00,170.00,219.00,273.00,281.00,73.00,332.00,64.00,243.00,251.00,268.00,154.00,49.00,477.00,287.00,239.00,164.00,403.00,167.00,248.00,359.00,279.00,433.00,226.00,298.00,270.00,187.00,171.00,296.00,251.00,221.00,273.00,276.00,224.00,201.00,106.00,178.00,208.00,219.00,103.00,293.00,258.00,246.00,285.00,276.00,185.00,239.00,213.00,290.00,139.00,170.00,244.00,251.00,253.00,154.00,210.00,161.00,244.00,257.00,427.00,488.00,428.00,231.00,218.00,79.00,238.00,295.00,83.00,494.00,288.00,147.00,130.00,160.00,197.00,281.00,551.00,367.00,468.00,469.00,205.00,131.00,281.00,110.00,373.00,294.00,270.00,210.00,284.00,117.00,224.00,175.00,110.00,386.00,490.00,307.00,200.00,208.00,184.00,148.00,264.00,271.00,162.00,229.00,86.00,333.00,358.00,116.00,252.00,189.00,224.00,202.00,414.00,368.00,248.00,68.00,195.00,237.00,297.00,135.00,123.00,265.00,163.00,47.00,485.00,445.00,19.00,104.00,178.00,215.00,245.00,176.00,180.00,283.00,167.00,87.00,376.00,53.00,204.00,273.00,146.00,78.00,234.00,127.00,45.00,239.00,338.00,355.00,138.00,314.00,96.00,225.00,170.00,167.00,171.00,191.00,120.00,452.00,447.00,378.00,239.00,129.00,172.00,171.00,79.00,85.00,218.00,190.00,81.00,263.00,203.00,295.00,250.00,147.00,178.00,193.00,187.00,174.00,293.00,122.00,190.00,149.00,248.00,134.00,174.00,305.00,324.00,168.00,182.00,215.00,234.00,120.00,214.00,210.00,251.00,121.00,82.00,159.00,304.00,97.00,318.00,172.00,459.00,247.00,55.00,260.00,387.00,226.00,98.00,175.00,258.00,295.00,260.00,201.00,278.00,80.00,245.00,281.00,288.00,120.00,261.00,263.00,341.00,230.00,201.00,129.00,153.00,411.00,156.00,172.00,203.00,292.00,190.00,254.00,116.00,147.00,265.00,119.00,352.00,229.00,89.00,153.00,337.00,130.00,466.00,359.00,170.00,171.00,269.00,87.00,392.00,135.00,205.00,180.00,423.00,533.00,263.00,236.00,264.00,251.00,393.00,401.00,169.00,208.00,238.00,291.00,241.00,200.00,111.00,226.00,138.00,114.00,247.00,247.00,129.00,107.00,223.00,333.00,323.00,134.00,322.00,109.00,304.00,248.00,254.00,290.00,224.00,364.00,252.00,174.00,100.00,103.00,283.00,179.00,181.00,254.00,215.00,179.00,315.00,135.00,330.00,508.00,471.00,97.00,289.00,88.00,64.00,106.00,114.00,170.00,261.00,252.00,268.00,259.00,190.00,77.00,214.00,235.00,261.00,112.00,480.00,174.00,194.00,195.00,154.00,273.00,317.00,218.00,153.00,212.00,223.00,217.00,136.00,212.00,135.00,228.00,383.00,263.00,228.00,268.00,312.00,164.00,291.00,222.00,195.00,213.00,217.00,193.00,286.00,126.00,154.00,272.00,102.00,191.00,136.00,363.00,146.00,198.00,262.00,151.00,162.00,202.00,108.00,40.00,125.00,215.00,185.00,196.00,193.00,157.00,264.00,22.00,370.00,210.00,263.00,396.00,502.00,170.00,261.00,216.00,108.00,161.00,207.00,332.00,412.00,113.00,193.00,233.00,165.00,165.00,108.00,130.00,208.00,225.00,335.00,9.00,199.00,231.00,101.00,195.00,278.00,152.00,215.00,213.00,84.00,80.00,255.00,234.00,172.00,553.00,442.00,251.00,170.00,294.00,131.00,379.00,81.00,264.00,283.00,101.00,207.00,214.00,127.00,115.00,214.00,99.00,326.00,190.00,231.00,97.00,556.00,392.00,158.00,216.00,223.00,199.00,263.00,99.00,518.00,339.00,281.00,195.00,238.00,253.00,281.00,282.00,52.00,414.00,148.00,310.00,73.00,252.00,232.00,412.00,174.00,320.00,198.00,543.00,192.00,245.00,130.00,166.00,145.00,254.00,327.00,83.00,175.00,180.00,95.00,247.00,130.00,240.00,133.00,160.00,224.00,120.00,159.00,122.00,73.00,114.00,232.00,186.00,50.00,241.00,102.00,236.00,258.00,87.00,454.00,272.00,258.00,110.00,307.00,376.00,198.00,75.00,151.00,409.00,263.00,113.00,266.00,301.00,188.00,296.00,169.00,275.00,296.00,250.00,295.00,37.00,294.00,245.00,318.00,160.00,35.00,313.00,83.00,125.00,179.00,272.00,162.00,64.00,102.00,121.00,321.00,145.00,244.00,67.00,257.00,154.00,147.00,221.00,201.00,322.00,82.00,195.00,88.00,274.00,254.00,164.00,264.00,143.00,311.00,136.00,227.00,244.00,322.00,206.00,49.00,255.00,128.00,316.00,282.00,135.00,186.00,298.00,141.00,230.00,119.00,282.00,351.00,373.00,113.00,36.00,185.00,459.00,175.00,75.00,252.00,78.00,58.00,99.00,100.00,159.00,134.00,238.00,288.00,150.00,243.00,133.00,269.00,258.00,189.00,193.00,374.00,161.00,190.00,235.00,312.00,64.00,287.00,250.00,251.00,119.00,261.00,73.00,207.00,510.00,77.00,242.00,201.00,283.00,180.00,190.00,241.00,206.00,250.00,344.00,221.00,433.00,175.00,267.00,147.00,527.00,464.00,525.00,254.00,243.00,66.00,134.00,295.00,125.00,170.00,164.00,317.00,350.00,333.00,192.00,234.00,420.00,340.00,182.00,273.00,321.00,476.00,154.00,217.00,47.00,462.00,218.00,82.00,269.00,137.00,500.00,271.00,262.00,114.00,181.00,754.00,1212.00,3110.00,5311.00,4489.00,2992.00,7344.00,8902.00,11277.00,16809.00,9208.00,10634.00,12481.00,2041.00,4246.00,2930.00,11066.00,13801.00,13123.00,9815.00,1207.00,3174.00,2948.00,7848.00,3850.00,3219.00,1731.00,6272.00,5206.00,4049.00,3163.00,4554.00,7089.00,5130.00,979.00,3131.00,4153.00,2245.00,2614.00,2779.00,1822.00,5444.00,5464.00,1187.00,3114.00,1796.00,1427.00,1736.00,642.00,970.00,1874.00,2413.00,1115.00,2616.00,1891.00,1296.00,1715.00,815.00,536.00,985.00,1478.00,1232.00,226.00,516.00,55.00,1787.00,3201.00,2713.00,1163.00,660.00,400.00,677.00,506.00,578.00,633.00,1218.00,1385.00,658.00,386.00,493.00,744.00,1035.00,939.00,1168.00,1095.00,648.00,998.00,1046.00,974.00,1208.00,1038.00,1066.00,696.00,350.00,348.00,706.00,617.00,410.00,294.00,226.00,279.00,843.00,886.00,971.00,510.00,358.00,541.00,89.00,299.00,213.00,458.00,653.00,583.00,428.00,547.00,298.00,383.00,590.00,819.00,686.00,740.00,345.00,93.00,272.00,262.00,216.00,300.00,281.00,293.00,508.00,297.00,243.00,478.00,34.00,534.00,480.00,530.00,156.00,289.00,389.00,441.00,150.00,217.00,223.00,55.00,261.00,106.00,479.00,81.00,208.00,81.00,21.00,500.00,411.00,238.00,273.00,272.00,256.00,145.00,173.00,315.00,220.00,279.00,181.00,69.00,281.00,142.00,132.00,196.00,165.00,234.00,124.00,353.00,256.00,170.00,160.00,93.00,68.00,225.00,361.00,113.00,88.00,77.00,198.00,178.00,339.00,195.00,178.00,504.00,94.00,511.00,51.00,106.00,340.00,70.00,181.00,96.00,233.00,187.00,206.00,285.00,157.00,181.00,143.00,228.00,472.00,442.00,199.00,87.00,243.00,216.00,107.00,310.00,95.00,92.00,222.00,209.00,373.00,303.00,69.00,196.00,180.00,127.00,129.00,103.00,302.00,264.00,262.00,266.00,199.00,202.00,192.00,280.00,31.00,200.00,234.00,122.00,445.00,89.00,329.00,288.00,139.00,160.00,217.00,550.00,215.00,257.00,217.00,279.00,30.00,314.00,234.00,138.00,196.00,264.00,251.00,83.00,154.00,82.00,324.00,208.00,32.00,373.00,224.00,174.00,124.00,329.00,207.00,96.00,218.00,0.00

Data set was created using SNPN-UWB board with DecaWave DWM1000 UWB radio module.



Solution

Our team implemented a sliding time window approach (window size = 3) to aggregate the raw CIR columns into averages, smoothing out erratic values. Following this, a variance filtering process was applied using a “knee point” analysis to identify and retain only the most informative data points. This process reduced the 1,016 original samples down to just a few highly relevant features that passed a 0.6 variance threshold, significantly streamlining the dataset without losing critical patterns.

CIR Columns Variance
CIR Columns Variance With Sliding Window Aggregation



The Lesson

Systematic dimensionality reduction and data aggregation are vital for transforming high-density raw signals into a machine-friendly format. Using statistical tools like variance plots to find “knee points” allows for a data-driven way to eliminate noise while preserving the features that contribute most to distinguishing signal patterns.





Technical Challenges & Solutions - Dealing with Multicollinearity


Issue

The feature set exhibited significant multicollinearity, where multiple variables were highly correlated and provided redundant information to the models. This redundancy introduces unnecessary complexity, which can cause machine learning models to prioritize noise over generalizable patterns, leading to poor performance when the model is exposed to new, unseen indoor environments.



Solution

Our team used Correlation Matrix Heatmaps to identify groups of features with strong correlations (thresholds of ±0.7). Next, Principal Component Analysis (PCA) was applied to these specific groups. Rather than just reducing dimensions, PCA was used here to transform the data into an orthogonal feature space, creating new, uncorrelated features that represent the maximum variance of the original data while eliminating the redundancy caused by multicollinearity.

Correlation Matrix Heatmap Between Features
Correlation Matrix Heatmap Between PCA Features



The Lesson

Mitigating multicollinearity is a critical step in improving model generalization. Using PCA to decorrelate features ensures that the model learns from unique information rather than being biased by redundant variables, making the final solution more robust and scalable for different indoor layouts.





Results

Classification Model Results
Regression Model Results



The classification and regression models used in this project were evaluated based on their ability to accurately distinguish between Line-of-Sight (LOS) and Non-Line-of-Sight (NLOS) conditions and predict signal ranges. Across both tasks, ensemble methods like XGBoost and Random Forest consistently outperformed other models due to their capacity to capture non-linear relationships within the UWB signal data.




Classification Model Results

The primary goal of classification was to categorize UWB Channel Impulse Response (CIR) measurements as either LOS or NLOS.


Top Performers: XGBoost achieved the highest accuracy at 91.67%, with Random Forest closely following at 91.25%. Both models demonstrated balanced precision and recall across both classes.


Baselines and Neural Networks: Logistic Regression, serving as the baseline, achieved a decent accuracy of 85.46%, suggesting some linear relationships in the data. The Multilayer Perceptron (MLP) had moderate performance at 80.99%, potentially limited by model complexity or the need for more fine-tuning.


Lowest Performer: K-Nearest Neighbour (KNN) performed the worst with 76.39% accuracy. Its poor performance was attributed to sensitivity to feature scaling and the “curse of dimensionality,” where high-dimensional data makes distance metrics less meaningful.



Regression Model Results

Regression models were tasked with predicting the continuous “measured range” for localization.


Top Performers: Random Forest and XGBoost were the most effective, producing the lowest Root Mean Squared Errors (RMSE) of 1.45 and 1.46, respectively. They also achieved the highest R² scores (0.6173 for Random Forest and 0.6128 for XGBoost), indicating they could explain over 60% of the variance in the target range.


Baselines and Neural Networks: Ridge Regression (baseline) reached an R² of 0.5220, but it was constrained by its linear nature and inability to model complex signal interactions. The MLP Regressor showed similar moderate results (R² of 0.4646), appearing to suffer from overfitting and a need for a larger dataset to compete with ensemble models.


Lowest Performer: KNN again performed the worst in regression, with a significantly higher RMSE of 1.96 and a low R² score of 0.2142. The results suggest that without more extensive feature engineering or scaling, KNN is not well-suited for this specific regression task.