how to understand the all_actions_scaled in ComputeOutput of rl_sdk.cpp

[furynerd]

hi

according to

rl_sar/src/rl_sar/library/core/rl_sdk/rl_sdk.cpp

Lines 256 to 272 in e384912

	void RL::ComputeOutput(const std::vector<float> &actions, std::vector<float> &output_dof_pos, std::vector<float> &output_dof_vel, std::vector<float> &output_dof_tau)
	{
	std::vector<float> actions_scaled = actions * this->params.Get<std::vector<float>>("action_scale");
	std::vector<float> pos_actions_scaled = actions_scaled;
	std::vector<float> vel_actions_scaled(actions.size(), 0.0f);
	for (int i : this->params.Get<std::vector<int>>("wheel_indices"))
	{
	pos_actions_scaled[i] = 0.0f;
	vel_actions_scaled[i] = actions_scaled[i];
	}
	std::vector<float> all_actions_scaled = pos_actions_scaled + vel_actions_scaled;
	output_dof_pos = pos_actions_scaled + this->params.Get<std::vector<float>>("default_dof_pos");
	output_dof_vel = vel_actions_scaled;
	output_dof_tau = this->params.Get<std::vector<float>>("rl_kp") * (all_actions_scaled + this->params.Get<std::vector<float>>("default_dof_pos") - this->obs.dof_pos) - this->params.Get<std::vector<float>>("rl_kd") * this->obs.dof_vel;
	output_dof_tau = clamp(output_dof_tau, -this->params.Get<std::vector<float>>("torque_limits"), this->params.Get<std::vector<float>>("torque_limits"));
	}

can anyone help me with the purpose of all_actions_scaled?

[1]
why all_actions_scaled is computed by adding elements of pos_actions_scaled + vel_actions_scaled?
it looks all_actions_scaled is indeed the same value with actions_scaled.

[2]
i think i understand pos_actions_scaled and vel_actions_scaled is needed for subsequent control purpose.
but all_actions_scaled is only used to calc output_dof_tau and which pushed into output_dof_tau_queue, no further utilization of all_actions_scaled , output_dof_tau or output_dof_tau_queue according to the searching from global code.

is all_actions_scaled a designed reservation or somewhere used for some purpose?

thanks very much