Tito Arevalo-Ramirez

Dr. Tito Arevalo-Ramirez

orcid

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
29 May 2025 in Robotics
Verified
Journal Article
Open access
PDF file
JMJosé Manuel Alcayaga Marín
José Manuel Alcayaga Marín
OMOswaldo Menéndez
Oswaldo Menéndez
MTMiguel Torres-Torriti
Miguel Torres-Torriti
Juan Pablo VásconezJuan Pablo Vásconez
Juan Pablo Vásconez
Tito Arevalo-RamirezTito Arevalo-Ramirez
Tito Arevalo-Ramirez
Alvaro Javier PradoAlvaro Javier Prado
Alvaro Javier Prado
Autonomous navigation in mining environments is challenged by complex wheel–terrain interaction, traction losses caused by slip dynamics, and sensor limitations. This paper investigates the effectiveness of Deep Reinforcement Learning (DRL) techniques for the trajectory tracking control of skid-steer mobile robots operating under terra-mechanical constraints. Four state-of-the-art DRL algorithms, i.e., Proximal Policy Optimization (PPO), Deep Deterministic Policy Gradient (DDPG), Twin Delayed DDPG (TD3), and Soft Actor–Critic (SAC), are selected to evaluate their ability to generate stable and adaptive control policies under varying environmental conditions. To address the inherent partial observability in real-world navigation, this study presents an original approach that integrates Long Short-Term Memory (LSTM) networks into DRL-based controllers. This allows control agents to retain and leverage temporal dependencies to infer unobservable system states. The developed agents were trained and tested in simulations and then assessed in field experiments under uneven terrain and dynamic model parameter changes that lead to traction losses in mining environments, targeting various trajectory tracking tasks, including lemniscate and squared-type reference trajectories. This contribution strengthens the robustness and adaptability of DRL agents by enabling better generalization of learned policies compared with their baseline counterparts, while also significantly improving trajectory tracking performance. In particular, LSTM-based controllers achieved reductions in tracking errors of 10%, 74%, 21%, and 37% for DDPG-LSTM, PPO-LSTM, TD3-LSTM, and SAC-LSTM, respectively, compared with their non-recurrent counterparts. Furthermore, DDPG-LSTM and TD3-LSTM reduced their control effort through the total variation in control input by 15% and 20% compared with their respective baseline controllers, respectively. Findings from this work provide valuable insights into the role of memory-augmented reinforcement learning for robust motion control in unstructured and high-uncertainty environments.

8 August 2023 in Forests
Verified
Journal Article
Open access
PDF file
MAMaria Eduarda Andrada
Maria Eduarda Andrada
DRDavid Jacob Russell
David Jacob Russell
Tito Arevalo-RamirezTito Arevalo-Ramirez
Tito Arevalo-Ramirez
WKWinnie Kuang
Winnie Kuang
GKGeorge Kantor
George Kantor
Francisco J. YandunFrancisco J. Yandun
Francisco J. Yandun
This paper presents a comprehensive forest mapping system using a customized drone payload equipped with Light Detection and Ranging (LiDAR), cameras, a Global Navigation Satellite System (GNSS), and Inertial Measurement Unit (IMU) sensors. The goal is to develop an efficient solution for collecting accurate forest data in dynamic environments and to highlight potential wildfire regions of interest to support precise forest management and conservation on the ground. Our paper provides a detailed description of the hardware and software components of the system, covering sensor synchronization, data acquisition, and processing. The overall system implements simultaneous localization and mapping (SLAM) techniques, particularly Fast LiDAR Inertial Odometry with Scan Context (FASTLIO-SC), and LiDAR Inertial Odometry Smoothing and Mapping (LIOSAM), for accurate odometry estimation and map generation. We also integrate a fuel mapping representation based on one of the models, used by the United States Secretary of Agriculture (USDA) to classify fire behavior, into the system using semantic segmentation, LiDAR camera registration, and odometry as inputs. Real-time representation of fuel properties is achieved through a lightweight map data structure at 4 Hz. The research results demonstrate the effectiveness and reliability of the proposed system and show that it can provide accurate forest data collection, accurate pose estimation, and comprehensive fuel mapping with precision values for the main segmented classes above 85%. Qualitative evaluations suggest the system’s capabilities and highlight its potential to improve forest management and conservation efforts. In summary, this study presents a versatile forest mapping system that provides accurate forest data for effective management.

15 September 2021 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
Tito Arevalo-RamirezTito Arevalo-Ramirez
Tito Arevalo-Ramirez
JGJavier Guevara
Javier Guevara
RRRobert Guaman Rivera
Robert Guaman Rivera
JVJuan Villacrés
Juan Villacrés
OMOswaldo Menéndez
Oswaldo Menéndez
FCFernando Auat Cheein
Fernando Auat Cheein
Bare-earth extraction in forested regions has been considered challenging because of the lack of ground point information. In these regions, vision systems cannot capture any information about ground points under the canopy. Thus, the challenge of generating a digital terrain model by cameras increases. Nevertheless, one might alleviate ground filtering using vegetation’s features (e.g., chlorophyll). In this regard, this article evaluated two machine-learning approaches [i.e., conditional random field (CRF), artificial neural network (ANN)] for generating digital terrain models when biophysical or biochemical features of vegetation are given. Terrain models were generated from multispectral image-based point clouds. A fivefold cross-validation methodology evaluated the CRF and ANN. The point clouds were retrieved from two study areas at different illumination and flight altitudes. Vegetation features were computed as vegetation indices from the multispectral point clouds. Results suggested that by using these indices, the classification of ground points could be enhanced. In particular, the vegetation indices that yielded the best outcomes were normalized difference vegetation index, green NDVI, and modified chlorophyll absorption reflectance index. Moreover, it was shown that CRF generates elevation models more smoothly than a triangular irregular network method. Thus, a CRF could be promising for classifying ground points in forested regions using geometric and vegetation features from a photogrammetric point cloud.

1 February 2021 in Biosystems Engineering
Verified
Journal Article
Closed access
Tito Arevalo-RamirezTito Arevalo-Ramirez
Tito Arevalo-Ramirez
ACAndrés Hernán Fuentes Castillo
Andrés Hernán Fuentes Castillo
PCPedro Sebastián Reszka Cabello
Pedro Sebastián Reszka Cabello
FCFernando Auat Cheein
Fernando Auat Cheein

23 February 2021 in IEEE Access
Verified
Journal Article
Open access
Tito Arevalo-RamirezTito Arevalo-Ramirez
Tito Arevalo-Ramirez
FCFernando Auat Cheein
Fernando Auat Cheein
The conditional random field is a suitable framework for contextual classification of two-dimensional (images) and three-dimensional (point clouds) data. This framework is based on probabilistic graphical models, an alternative representation of a conditional probability distribution over random variables. In general, a graphical model encodes probabilistic relationships between random variables by their edges. However, graph structure (set of edges) is not always known in advance. In particular, if we consider each point of a point cloud as a graph node, we might not have information about nodes interaction (i.e. graph structure is unknown). Given that there is no agreement about what structure to use for point cloud contextual classification, we focused on determining a suitable graph structure by comparing the performance of four different graph structures and twenty feature sets. All experiments were performed in urban environments. The quantitative errors (type I, type II and total error) and classification accuracy were the metrics we used to evaluate the performance of Conditional Random Fields. Results suggest that optimal neighbor and 3D Delaunay structures achieved best classification performance. Theses structures combined with point-slope, curvature, and segment height node based features performed a classification accuracy error of 93.5% and the best type I error, 1.56%.

26 June 2024 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
Tito Arevalo-RamirezTito Arevalo-Ramirez
Tito Arevalo-Ramirez
AAAnali Alfaro
Anali Alfaro
JSJose M. Saavedra
Jose M. Saavedra
MRMatias Recabarren
Matias Recabarren
MPMauricio Ponce-Donoso
Mauricio Ponce-Donoso
JDJose Delpiano
Jose Delpiano
Deep learning has gained popularity in recent years for reconstructing hyperspectral and multispectral images, offering cost-effective solutions and promising results. Research on hyperspectral image reconstruction feeds deep learning models with images at specific wavelengths and outputs images in other spectral bands. Although encouraging results of previous works, it should be determined to what extent the reconstructed information can lead to an advantage over the captured images. In this context, the present work inspects whether or not reconstructed spectral images add relevant information to segmentation networks for improving urban tree identification. Specifically, we generate red-edge (ReD) and near-infrared (NIR) images from RGB images using a conditional Generative Adversarial Network (cGAN). The training and validation are carried out with 5770 multispectral images obtained after a custom data augmentation process using an urban hyperspectral dataset. The testing outcomes reveal that ReD and NIR can be generated with an average Structural Similarity Index Measure of 0.93 and 0.88, respectively. Next, the cGAN generates ReD and NIR information of two RGB-based urban tree datasets (i.e., Jekyll, 3949 samples, and Arbocensus, 317 samples). Subsequently, DeepLabV3 and SegFormer segmentation networks are trained, validated, and tested using RGB, RGB+ReD, and RGB+NIR images from Jekyll and Arbocensus datasets. The experiments show that reconstructed multispectral images might not add information to segmentation networks that enhance their performance. Specifically, the p-values from a T-test show no significant difference between the performance of segmentation networks.