Table 2.

Comparison of various conventional cooperative control.

Control algorithms	Advantages	Challenges	References
1. HIC	Tremendous force tracking features and high position flexibility.	Improper adjustments can create system instability.	Li K. et al., 2020
2. Navel AVIC	Dynamic contact force detection in an uncertain environment, unknown environmental stiffness, and dynamic environmental location evaluation can also be applied to slope surfaces or other complex areas. It can yield a hugely convincing solution to the contact operation.	Complex.	Duan et al., 2018
3. PD-based position control	Feasible, and it can provide outstanding flexibility and positional accuracy.	It only works well for linear systems, and a large time delay process is required for parameter adjustments.	Zeng et al., 2019
4. Radial basis function (RBF) neural network compensation control	The uncertainty part of computed torque, as well as the disturbance of rehabilitation, can be eliminated, which will optimize tracking ability.	Only the nonlinear system can be effectively trained without a time-consuming training process and dynamical model evaluation.	Yu et al., 2011; Yin et al., 2015; Tao et al., 2016
5. Compensation methods and reaction torque observer (ROB)	The technique can satisfy and meet the requirements of patient-cooperative control, it is easy to design, and has good accuracy.	-	Van Tran et al., 2015a,b
6. Adaptive impedance monitoring with BP neural network	The patient can effectively participate in rehabilitation training.	-	Chen et al., 2020
7. Under actuated energy shaping technique	The approach can overcome this issue and yield task-invariant, trajectory-free control laws that can be suitable for the various DLAs.	Complex.	Lv and Gregg, 2017; Lv et al., 2018
8. Fuzzy PD position controller	It is stable and smooth, minimizes deviation, and optimizes the system's dynamic performance.	Rarely are fuzzy inference systems used directly in the control loop and sometimes low control.	Bai et al., 2017
9. Trajectory deformation algorithm (TDA)	TDA can promote the trajectory of rehabilitation training more naturally and smoother, enhance robot compliance, and protect the user from secondary injury.	The tracking trajectory's smoothness is significantly worse than that of the desired trajectory.	Zhou et al., 2021
10. Fuzzy NN	Excellent at handling uncertain information.	It cannot handle the vague the message, slow processing speed, low accuracy, and massive linear and angular velocities that lead to oscillation problems.	Bai et al., 2017