IEEE Robotics & Automation Magazine - March 2021 - 21

such as the SHUYU robot [Tsinghua University, China,
Figure 3(b)] [13], which locates human faces through computer vision and takes the temperature using a noncontact
infrared thermometer.
The main challenges for improved outcomes regard the
accuracy and robustness of the sensors. For instance, thermal
cameras may fail in detecting the correct temperature when
particular conditions are encountered, such as when sweat or
a mask covers the face of the subject. Increasing the environment and context awareness of the robot would be beneficial
to tackle these difficulties so that similar errors could be compensated for with the aid of additional sensors and computational processing.
Robots for Diagnosis
Bio Sampling and Image-Guided Diagnosis
Depending on the disease, a host of samples may need to be
collected, such as blood or stool samples in the case of nonairborne diseases like Ebola and cholera, or saliva, oral, or nasal
swab samples in the case of airborne diseases like COVID-19.
Conventional testing methods typically require interaction
between a potentially infected patient and medical workers,
thus representing occasions for the potential spreading of the
virus. Additional issues may also stem from the handling of
collected samples prior to, during, or after testing. Hence,
robotic solutions for collecting, handling, testing, and disposing of these samples may allow for a valuable reduction in
transmission of and exposure to a disease.
Telerobots with manipulation capabilities are able to achieve
physical human-robot interaction, which is not feasible with
conventional telemedicine solutions. As an example, the Chinese Academy of Sciences developed steerable telerobots for
throat swab sampling of coronavirus tests [Figure 4(a)] [14].
During traditional throat swab sampling, health-care staff is in
close contact with patients, which poses a high risk of cross
infection. In addition, health-care workers' operating skills
affect the accuracy and quality of swab results. To overcome
those limitations, health-care workers can teleoperate the robot
with both haptic and visual feedback from the high-definition
3D anatomical view of binocular endoscopes.

(a)

(b)

Figure 3. Robots for screening. (a) A robot used for temperature
monitoring at a hospital in Shenyang, China. (b) The SHUYU
robot developed by Tsinghua University, China, allows for rapid
drive-through temperature screening [13].

A further step was taken by a joint team from Lifeline
Robotics and the University of Southern Denmark, who
developed the first fully automatic throat swab robot (Figure 1). Another work presented a portable robot [Figure 4(b)]
for needle placement to draw blood or deliver fluids through
image-guided autonomous operation [15]. Multimodal image
sequences (both ultrasound and near-infrared optical imaging) were decoded by predictions from a series of deep convolutional neural networks to guide the real-time actuation of
the robotic cannulation process.
Another important method for diagnosis, especially during the COVID-19 pandemic, is ultrasonic examination,
which is well suited for monitoring the condition of the
lungs, unlike the computerized tomography scan, which
causes radiation and is not in real time. Tsinghua University
evaluated a force-controlled ultrasound robot [Figure 4(c)]
that fuses cross-modal sensory information from ultrasound and force measurements for remote diagnosis to
minimize the contact between health-care staff and patients.
The University of Maryland developed a semi-autonomous
system for hemorrhage detection using robotic ultrasound
[16] and explored using the system for COVID-19 lung
imaging [Figure 4(d)]. A similar study was conducted to
evaluate the feasibility of a remote-robot-assisted ultrasound
system in examining patients with COVID-19 [17].
Tool Camera

Robot

Ultrasound
Probe

Force Sensor

(a)

(b)

(c)

Ultrasound
Phantom
(d)

Figure 4. Robots for sampling and diagnosis. (a) A steerable telerobot for throat swabbing from the Chinese Academy of Sciences [14].
(b) A portable robot exploiting deep learning for blood testing at Rutgers University, United States [15]. (c) A robotic platform for lung
ultrasound at Tsinghua University, China. (d) A robotic system for remote trauma assessment at the University of Maryland, United
States [16].

MARCH 2021

*

IEEE ROBOTICS & AUTOMATION MAGAZINE

*

21
IEEE Robotics & Automation Magazine - March 2021

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - March 2021
