IR-MCL: Implicit Representation-Based Online Global Localization

Kuang, Haofei; Chen, Xieyuanli; Guadagnino, Tiziano; Zimmerman, Nicky; Behley, Jens; Stachniss, Cyrill

doi:10.1109/lra.2023.3239318

Cited by 18 publications

(6 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The last was testing the AMCL+CNN algorithm, which is the method proposed in this paper. In addition, the test results were also compared with that of earlier study that proposed IR-MCL [26]. Several testing were then compared to determine the average error level that occurred.…”

Section: Resultsmentioning

confidence: 99%

“…Compared with these methods, the proposed method, AMCL+CNN, yielded the smallest position error, which was 0.0202 m (the average of straight and turning scenarios). The NMCL algorithm in testing produced a position error value of 0.1369 m, while the IR-MCL produced a position error of 0.0687 [26]. However, the value of the heading error from AMCL+CNN is the biggest heading error, so it still has the potential for improvement.…”

Section: Testing With the Amcl+cnnmentioning

confidence: 99%

“…Additionally, the results of conducted research [26] were used. In this testing, the estimated position of the robot was tested using the Monte Carlo localization + exploiting text spotting (NMCL) and implicit representation-based MCL (IR-MCL).…”

Section: Testing With the Amcl+cnnmentioning

confidence: 99%

“…The accuracy of reading AMCL can also be increased using implicit representation-based Monte Carlo localization (IR-MCL) [26]. This technique performs position estimation based on 2D LiDAR data.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Peningkatan Akurasi Adaptive Monte Carlo Localization Menggunakan Convolutional Neural Network

Riza Agung Firmansyah,

Tri Arief Sardjono,

Ronny Mardiyanto

2023

JNTETI

View full text Add to dashboard Cite

Makalah ini menjelaskan peningkatan akurasi sistem lokalisasi adaptive Monte Carlo localization (AMCL) pada robot menggunakan convolutional neural network (CNN). Sistem lokalisasi pada sebuah robot merupakan proses pengenalan posisi suatu robot dalam lingkungan kerjanya. Sistem ini penting karena memungkinkan robot untuk melakukan navigasi dan pemetaan (mapping) secara efisien dan akurat. Tanpa lokalisasi yang tepat, robot tidak dapat beroperasi dengan efektif dan dapat mengalami masalah, seperti kehilangan arah atau menabrak objek. AMCL merupakan sistem lokalisasi yang populer dan banyak digunakan dalam robot. Metode ini menggunakan perubahan posisi robot dan perubahan pembacaan sensor light detection and ranging (LiDAR) sebagai masukannya. Pembacaan perubahan posisi robot rentan terhadap kesalahan akibat slip maupun deformasi roda. Ketidaktepatan pembacaan perubahan posisi robot menyebabkan prediksi posisi robot oleh AMCL menjadi kurang tepat, sehingga diperlukan sebuah perbaikan. Pada makalah ini, kebaruan yang diangkat adalah memberikan nilai kompensasi dari hasil AMCL, sehingga error-nya menjadi kecil. Nilai kompensasi tersebut didapatkan dari hasil pelatihan CNN, sehingga metode yang diusulkan tersebut dapat disebut AMCL+CNN. Masukan yang diberikan pada CNN adalah perubahan nilai odometry roda dan perubahan pembacaan jarak oleh sensor LiDAR. Keluaran CNN tersebut dibandingkan dengan data target berupa posisi aktual robot dari hasil pengamatan. Pelatihan jaringan dilakukan sebanyak 200 epoch untuk mendapatkan validation loss terkecil. Pengujian dilakukan pada sebuah robot yang telah dilengkapi dengan robot operating system (ROS). Dataset pelatihan dan pengujian yang diujikan berasal dari data rosbag saat robot berjalan di area pengujian. Hasil yang didapatkan algoritma AMCL+CNN memiliki error yang lebih kecil dibandingkan AMCL dan Monte Carlo localization (MCL) reguler pada skenario lurus dan belok. Hasil yang diperoleh juga lebih baik pada metrik error posisi saat dibandingkan dengan beberapa metode pembanding lain.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Testing With the Amcl+cnnmentioning

confidence: 99%

Section: Testing With the Amcl+cnnmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Peningkatan Akurasi Adaptive Monte Carlo Localization Menggunakan Convolutional Neural Network

Riza Agung Firmansyah,

Tri Arief Sardjono,

Ronny Mardiyanto

2023

JNTETI

View full text Add to dashboard Cite

show abstract

“…The method performs comparably to a baseline system using a conventional LiDAR-based prior map. Haofei Kuang et al [ 49 ] proposed a global localization method for mobile robots using 2D LiDAR. In this method, the neural occupancy field uses a neural network to implicitly represent the scene, and 2D LiDAR scans of arbitrary robot poses are synthesized by using a pre-trained network.…”

Section: Overview Of Single Sensor Sensing Technologiesmentioning

confidence: 99%

A Review of Sensing Technologies for Indoor Autonomous Mobile Robots

Liu,

Wang,

Xie

et al. 2024

Sensors

View full text Add to dashboard Cite

As a fundamental issue in robotics academia and industry, indoor autonomous mobile robots (AMRs) have been extensively studied. For AMRs, it is crucial to obtain information about their working environment and themselves, which can be realized through sensors and the extraction of corresponding information from the measurements of these sensors. The application of sensing technologies can enable mobile robots to perform localization, mapping, target or obstacle recognition, and motion tasks, etc. This paper reviews sensing technologies for autonomous mobile robots in indoor scenes. The benefits and potential problems of using a single sensor in application are analyzed and compared, and the basic principles and popular algorithms used in processing these sensor data are introduced. In addition, some mainstream technologies of multi-sensor fusion are introduced. Finally, this paper discusses the future development trends in the sensing technology for autonomous mobile robots in indoor scenes, as well as the challenges in the practical application environments.

show abstract