Page 82 - Read Online

Page 77 Li et al. Intell Robot 2021;1(1):58-83 I http://dx.doi.org/10.20517/ir.2021.08

Figure 15. The simulation experiment in the case that the leader robot 3 is broken down [26] .

ance in the 3-D underwater environment was proposed by Ding et al. [115] . The bio-inspired neural network
helps the leader UUV decide the transform of the formation when encountering obstacle fields to avoid the
obstacles for all UUVs and meanwhile sustain on the desired trajectory. However, complex environmental
disturbances such as multi-obstacles are not thoroughly considered in this paper.

5. CHALLENGES AND FUTURE WORKS
Although there have been many studies of bio-inspired intelligence with applications to robotics and remark-
able achievements have been accomplished, there are still several challenges that would be further investigated
as future works.
• Manyexistingapproachesassumedtheenvironmentisstaticandwithoutanyuncertainties(e.g., somerobot
breakdown), disturbance (e.g., wind for surface robots, ocean/river current for underwater robots), and
noise (system and measurement noise). However, it is a big challenge for collision-free robot navigation in
complexchangingenvironmentswithmanymovingrobots/targetsandsubjecttouncertainties,disturbance,
and noise.
• Communication issue has always been an essential research area in robotics. It is important to build a
stable communication network in multi-robotic systems to ensure the updating of neural activity in the
bio-inspired neural network. However, most studies on the cooperation of multi-robotic systems did not
consider the communication issue, where the communication is normally noisy and with time delay. Many
approaches did not consider the optimal performance with multi-objectives (e.g., short total distance, com-
pletion time, energy, smoothness of the robot paths). Communication and multi-objectives optimization
could be a potential research direction in the future.
• Most conventional aerial robot navigation cannot act properly due to the limitations of communication
and perception ability of sensors in complex environments. The complexity of the aerial robot makes the
controllers are hard to design to achieve overall good performance. Though real-time collision-free nav-
igation and control of mobile robots, surface robots, and underwater robots have been studied for many
years, there is a lack of research for aerial robots based on bio-inspired neurodynamics models. The future
research is to incorporate bio-inspired neurodynamics models with other useful algorithms for aerial robot
navigation.