Hebert Azevedo-Sa
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Hebert Azevedo-Sa has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
- Dawn M. Tilbury
University of Michigan, Ann Arbor, MI, USA
4 shared publications - Lionel Peter Robert
University of Michigan, Ann Arbor, USA
4 shared publications - Suresh Kumaar Jayaraman
Department of Mechanical Engineering, University of Michigan, Ann Arbor, MI, USA
3 shared publications - Xi Jessie Yang
University of Michigan, Ann Arbor, USA
3 shared publications Armando Morado Ferreira
Brazilian Army’s Command and General Staff College (ECEME)
2 shared publications- Connor Esterwood
School of Information, University of Michigan, Ann Arbor, MI, USA
2 shared publications
Latest Publications
- Open Access
- 1 January 2025
- Published by FapUNIFESP (SciELO) in Boletim de Ciências Geodésicas
- Vol. 31
Single-point positioning performance using ionospheric corrections predicted by the ED-ConvLSTM-ND neural network
Predicted Global Ionospheric Maps (GIMs) are widely used in single-frequency Global Navigation Satellite Systems (GNSS) applications to correct ionospheric delays and enhance receiver positioning accuracy. In this work, we employed the Encoder-Decoder Convolutional Long-Short Term Memory for the Next Day (ED-ConvLSTM-ND) recurrent neural network to predict next-day GIMs based on GIMs from previous days. Although the model was originally evaluated in the ionospheric map domain-by comparing the Vertical Total Electron Content (VTEC) of the predicted maps with post-processed GIMs-such assessments may not fully capture the impact on positioning performance. To address this, we evaluated whether the Ionosphere Map Exchange (IONEX) files generated by the ED-ConvLSTM-ND network improve the accuracy of single-frequency single-point positioning (SF-SPP). We compared time series generated using various ionospheric correction approaches: the Klobuchar model (BRDC), the Center for Orbit Determination in Europe (CODE) 1-day predicted GIM (C1PG), the ED-ConvLSTM-ND GIM, and the final CODE GIM (CODG), adopted as the reference ionosphere. Data from three continuously operating stations in Brazil, collected during 2015, were used. The ED-ConvLSTM-ND GIM achieved a mean absolute error (MAE) 7.7% lower than that of C1PG. ED-ConvLSTM-ND predicted GIMs obtained MAE at least 0.76 m better than Klobuchar in every station, indicating that neural networks could improve real-time GNSS positioning.
- 1 January 2025
- Published by FapUNIFESP (SciELO) in Boletim de Ciências Geodésicas
- Vol. 31
Open AccessSingle-point positioning performance using ionospheric corrections predicted by the ED-ConvLSTM-ND neural network
Predicted Global Ionospheric Maps (GIMs) are widely used in single-frequency Global Navigation Satellite Systems (GNSS) applications to correct ionospheric delays and enhance receiver positioning accuracy. In this work, we employed the Encoder-Decoder Convolutional Long-Short Term Memory for the Next Day (ED-ConvLSTM-ND) recurrent neural network to predict next-day GIMs based on GIMs from previous days. Although the model was originally evaluated in the ionospheric map domain-by comparing the Vertical Total Electron Content (VTEC) of the predicted maps with post-processed GIMs-such assessments may not fully capture the impact on positioning performance. To address this, we evaluated whether the Ionosphere Map Exchange (IONEX) files generated by the ED-ConvLSTM-ND network improve the accuracy of single-frequency single-point positioning (SF-SPP). We compared time series generated using various ionospheric correction approaches: the Klobuchar model (BRDC), the Center for Orbit Determination in Europe (CODE) 1-day predicted GIM (C1PG), the ED-ConvLSTM-ND GIM, and the final CODE GIM (CODG), adopted as the reference ionosphere. Data from three continuously operating stations in Brazil, collected during 2015, were used. The ED-ConvLSTM-ND GIM achieved a mean absolute error (MAE) 7.7% lower than that of C1PG. ED-ConvLSTM-ND predicted GIMs obtained MAE at least 0.76 m better than Klobuchar in every station, indicating that neural networks could improve real-time GNSS positioning.
- Open Access
- 12 September 2022
- Published by Springer Nature in Scientific Reports
- Vol. 12 (1) , 1-15
Heterogeneous human–robot task allocation based on artificial trust
Effective human–robot collaboration requires the appropriate allocation of indivisible tasks between humans and robots. A task allocation method that appropriately makes use of the unique capabilities of each agent (either a human or a robot) can improve team performance. This paper presents a novel task allocation method for heterogeneous human–robot teams based on artificial trust from a robot that can learn agent capabilities over time and allocate both existing and novel tasks. Tasks are allocated to the agent that maximizes the expected total reward. The expected total reward incorporates trust in the agent to successfully execute the task as well as the task reward and cost associated with using that agent for that task. Trust in an agent is computed from an artificial trust model, where trust is assessed along a capability dimension by comparing the belief in agent capabilities with the task requirements. An agent’s capabilities are represented by a belief distribution and learned using stochastic task outcomes. Our task allocation method was simulated for a human–robot dyad. The team total reward of our artificial trust-based task allocation method outperforms other methods both when the human’s capabilities are initially unknown and when the human’s capabilities belief distribution has converged to the human’s actual capabilities. Our task allocation method enables human–robot teams to maximize their joint performance.
- 12 September 2022
- Published by Springer Nature in Scientific Reports
- Vol. 12 (1) , 1-15
Open AccessHeterogeneous human–robot task allocation based on artificial trust
Effective human–robot collaboration requires the appropriate allocation of indivisible tasks between humans and robots. A task allocation method that appropriately makes use of the unique capabilities of each agent (either a human or a robot) can improve team performance. This paper presents a novel task allocation method for heterogeneous human–robot teams based on artificial trust from a robot that can learn agent capabilities over time and allocate both existing and novel tasks. Tasks are allocated to the agent that maximizes the expected total reward. The expected total reward incorporates trust in the agent to successfully execute the task as well as the task reward and cost associated with using that agent for that task. Trust in an agent is computed from an artificial trust model, where trust is assessed along a capability dimension by comparing the belief in agent capabilities with the task requirements. An agent’s capabilities are represented by a belief distribution and learned using stochastic task outcomes. Our task allocation method was simulated for a human–robot dyad. The team total reward of our artificial trust-based task allocation method outperforms other methods both when the human’s capabilities are initially unknown and when the human’s capabilities belief distribution has converged to the human’s actual capabilities. Our task allocation method enables human–robot teams to maximize their joint performance.
- Open Access
- 9 June 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 6 (3) , 5913-5920
A Unified Bi-Directional Model for Natural and Artificial Trust in Human–Robot Collaboration
We introduce a novel capabilities-based bi-directional multi-task trust model that can be used for trust prediction from either a human or a robotic trustor agent. Tasks are represented in terms of their capability requirements, while trustee agents are characterized by their individual capabilities. Trustee agents' capabilities are not deterministic; they are represented by belief distributions. For each task to be executed, a higher level of trust is assigned to trustee agents who have demonstrated that their capabilities exceed the task's requirements. We report results of an online experiment with 284 participants, revealing that our model outperforms existing models for multi-task trust prediction from a human trustor. We also present simulations of the model for determining trust from a robotic trustor. Our model is useful for control authority allocation applications that involve human-robot teams.
- 9 June 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 6 (3) , 5913-5920
Open AccessA Unified Bi-Directional Model for Natural and Artificial Trust in Human–Robot Collaboration
We introduce a novel capabilities-based bi-directional multi-task trust model that can be used for trust prediction from either a human or a robotic trustor agent. Tasks are represented in terms of their capability requirements, while trustee agents are characterized by their individual capabilities. Trustee agents' capabilities are not deterministic; they are represented by belief distributions. For each task to be executed, a higher level of trust is assigned to trustee agents who have demonstrated that their capabilities exceed the task's requirements. We report results of an online experiment with 284 participants, revealing that our model outperforms existing models for multi-task trust prediction from a human trustor. We also present simulations of the model for determining trust from a robotic trustor. Our model is useful for control authority allocation applications that involve human-robot teams.
- Open Access
- 24 September 2020
- Published by Springer Nature in International Journal of Social Robotics
- Vol. 13 (8) , 1911-1927
Real-Time Estimation of Drivers’ Trust in Automated Driving Systems
Trust miscalibration issues, represented by undertrust and overtrust, hinder the interaction between drivers and self-driving vehicles. A modern challenge for automotive engineers is to avoid these trust miscalibration issues through the development of techniques for measuring drivers’ trust in the automated driving system during real-time applications execution. One possible approach for measuring trust is through modeling its dynamics and subsequently applying classical state estimation methods. This paper proposes a framework for modeling the dynamics of drivers’ trust in automated driving systems and also for estimating these varying trust levels. The estimation method integrates sensed behaviors (from the driver) through a Kalman filter-based approach. The sensed behaviors include eye-tracking signals, the usage time of the system, and drivers’ performance on a non-driving-related task. We conducted a study ( $$n=80$$ ) with a simulated SAE level 3 automated driving system, and analyzed the factors that impacted drivers’ trust in the system. Data from the user study were also used for the identification of the trust model parameters. Results show that the proposed approach was successful in computing trust estimates over successive interactions between the driver and the automated driving system. These results encourage the use of strategies for modeling and estimating trust in automated driving systems. Such trust measurement technique paves a path for the design of trust-aware automated driving systems capable of changing their behaviors to control drivers’ trust levels to mitigate both undertrust and overtrust.- 24 September 2020
- Published by Springer Nature in International Journal of Social Robotics
- Vol. 13 (8) , 1911-1927
Open AccessReal-Time Estimation of Drivers’ Trust in Automated Driving Systems
Trust miscalibration issues, represented by undertrust and overtrust, hinder the interaction between drivers and self-driving vehicles. A modern challenge for automotive engineers is to avoid these trust miscalibration issues through the development of techniques for measuring drivers’ trust in the automated driving system during real-time applications execution. One possible approach for measuring trust is through modeling its dynamics and subsequently applying classical state estimation methods. This paper proposes a framework for modeling the dynamics of drivers’ trust in automated driving systems and also for estimating these varying trust levels. The estimation method integrates sensed behaviors (from the driver) through a Kalman filter-based approach. The sensed behaviors include eye-tracking signals, the usage time of the system, and drivers’ performance on a non-driving-related task. We conducted a study ( $$n=80$$ ) with a simulated SAE level 3 automated driving system, and analyzed the factors that impacted drivers’ trust in the system. Data from the user study were also used for the identification of the trust model parameters. Results show that the proposed approach was successful in computing trust estimates over successive interactions between the driver and the automated driving system. These results encourage the use of strategies for modeling and estimating trust in automated driving systems. Such trust measurement technique paves a path for the design of trust-aware automated driving systems capable of changing their behaviors to control drivers’ trust levels to mitigate both undertrust and overtrust. - Open Access
- 22 September 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 5 (4) , 6908-6915
Context-Adaptive Management of Drivers’ Trust in Automated Vehicles
Automated vehicles (AVs) that intelligently interact with drivers must build a trustworthy relationship with them. A calibrated level of trust is fundamental for the AV and the driver to collaborate as a team. Techniques that allow AVs to perceive drivers’ trust from drivers’ behaviors and react accordingly are, therefore, needed for context-aware systems designed to avoid trust miscalibrations. This letter proposes a framework for the management of drivers’ trust in AVs. The framework is based on the identification of trust miscalibrations (when drivers’ undertrust or overtrust the AV) and on the activation of different communication styles to encourage or warn the driver when deemed necessary. Our results show that the management framework is effective, increasing (decreasing) trust of undertrusting (overtrusting) drivers, and reducing the average trust miscalibration time periods by approximately 40%. The framework is applicable for the design of SAE Level 3 automated driving systems and has the potential to improve the performance and safety of driver–AV teams.
- 22 September 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 5 (4) , 6908-6915
Open AccessContext-Adaptive Management of Drivers’ Trust in Automated Vehicles
Automated vehicles (AVs) that intelligently interact with drivers must build a trustworthy relationship with them. A calibrated level of trust is fundamental for the AV and the driver to collaborate as a team. Techniques that allow AVs to perceive drivers’ trust from drivers’ behaviors and react accordingly are, therefore, needed for context-aware systems designed to avoid trust miscalibrations. This letter proposes a framework for the management of drivers’ trust in AVs. The framework is based on the identification of trust miscalibrations (when drivers’ undertrust or overtrust the AV) and on the activation of different communication styles to encourage or warn the driver when deemed necessary. Our results show that the management framework is effective, increasing (decreasing) trust of undertrusting (overtrusting) drivers, and reducing the average trust miscalibration time periods by approximately 40%. The framework is applicable for the design of SAE Level 3 automated driving systems and has the potential to improve the performance and safety of driver–AV teams.
- Open Access
- 23 March 2020
- Published by Association for Computing Machinery (ACM)
- p. 113-115
Comparing the Effects of False Alarms and Misses on Humans' Trust in (Semi)Autonomous Vehicles
Trust in automated driving systems is crucial for effective driver-(semi)autonomous vehicles interaction. Drivers that do not trust the system appropriately are not able to leverage its benefits. This study presents a mixed design user experiment where participants conducted a non-driving task while traveling in a simulated semi-autonomous vehicle with forward collision alarm and emergency braking functions. Occasionally, the system missed obstacles or provided false alarms. We varied these system error types as well as road shapes, and measured the effects of these variations on trust development. Results reveal that misses are more harmful to trust development than false alarms, and that these effects are strengthened by operation on risky roads. Our findings provide additional insight into the development of trust in automated driving systems, and are useful for the design of such technologies.
- 23 March 2020
- Published by Association for Computing Machinery (ACM)
- p. 113-115
Open AccessComparing the Effects of False Alarms and Misses on Humans' Trust in (Semi)Autonomous Vehicles
Trust in automated driving systems is crucial for effective driver-(semi)autonomous vehicles interaction. Drivers that do not trust the system appropriately are not able to leverage its benefits. This study presents a mixed design user experiment where participants conducted a non-driving task while traveling in a simulated semi-autonomous vehicle with forward collision alarm and emergency braking functions. Occasionally, the system missed obstacles or provided false alarms. We varied these system error types as well as road shapes, and measured the effects of these variations on trust development. Results reveal that misses are more harmful to trust development than false alarms, and that these effects are strengthened by operation on risky roads. Our findings provide additional insight into the development of trust in automated driving systems, and are useful for the design of such technologies.
- 22 September 2015
- Published by SAE International in SAE International Journal of Advances and Current Practices in Mobility
- Vol. 1
Fuzzy Shared Semi-Autonomous Control System For Military Vehicles
Semi-autonomous control systems applied to automobiles are Advanced Driver Assistance Systems (ADAS) that have gained importance from similar devices with applications in robotics. The control sharing between humans and automatic controllers is the main characteristic of these systems, and can be accomplished through various different manners. However, the use of Artificial Intelligence (AI) techniques for this purpose remains unexplored. In this paper we propose the design of a semi-autonomous control system applied to military vehicles through the use of Fuzzy Inference Systems for the definition of the controller intervention level. Simulations of a vehicle being operated in highly dangerous situations, represented by the existence of hostile military threats or by unexpected maneuvers that could put the stability of the car at risk were performed. The control system’s level of intervention during the simulations was observed, and we could realize the increase of this variable according to the level of threat that the car was exposed to. The application of the proposed system results in safer operation of the vehicle, which shall be controlled with greater influence of the automatic controller when in greater danger. We present a critical analysis of these results and new directions for the future of this work.
- 22 September 2015
- Published by SAE International in SAE International Journal of Advances and Current Practices in Mobility
- Vol. 1
Fuzzy Shared Semi-Autonomous Control System For Military Vehicles
Semi-autonomous control systems applied to automobiles are Advanced Driver Assistance Systems (ADAS) that have gained importance from similar devices with applications in robotics. The control sharing between humans and automatic controllers is the main characteristic of these systems, and can be accomplished through various different manners. However, the use of Artificial Intelligence (AI) techniques for this purpose remains unexplored. In this paper we propose the design of a semi-autonomous control system applied to military vehicles through the use of Fuzzy Inference Systems for the definition of the controller intervention level. Simulations of a vehicle being operated in highly dangerous situations, represented by the existence of hostile military threats or by unexpected maneuvers that could put the stability of the car at risk were performed. The control system’s level of intervention during the simulations was observed, and we could realize the increase of this variable according to the level of threat that the car was exposed to. The application of the proposed system results in safer operation of the vehicle, which shall be controlled with greater influence of the automatic controller when in greater danger. We present a critical analysis of these results and new directions for the future of this work.
- 1 March 2012
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 21643342,p. 934-938
A field assessment of HF/VHF wire antenna impedance changes in rain forests
This work addresses the effects of vegetation coupling to HF/VHF wire antennas used in dense woods. A preliminary theoretical evaluation was carried out for wire dipoles, in which the electrical parameters of a thick forest were considered in the antenna resistance and reactance equations, instead of the vacuum properties. Then, impedance measurements of a few wire antennas for tactical radio links were performed in a vegetated site of Manaus, in the Brazilian Amazon. Comparing the open area to the forest results, moderate frequency shifts to the left were observed in the latter, as expected from the predictions.
- 1 March 2012
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 21643342,p. 934-938
A field assessment of HF/VHF wire antenna impedance changes in rain forests
This work addresses the effects of vegetation coupling to HF/VHF wire antennas used in dense woods. A preliminary theoretical evaluation was carried out for wire dipoles, in which the electrical parameters of a thick forest were considered in the antenna resistance and reactance equations, instead of the vacuum properties. Then, impedance measurements of a few wire antennas for tactical radio links were performed in a vegetated site of Manaus, in the Brazilian Amazon. Comparing the open area to the forest results, moderate frequency shifts to the left were observed in the latter, as expected from the predictions.