Comment out unnecessary gradient hessian estimation steps, hessian is anyways unused (c_hess=0)

Author: addat10

flocking_analysis/simulation_scripts/FlockingVehicleAgent.m

@@ -128,35 +128,36 @@
             % This function estimates the gradient and the hessian of the 
             % field at the current agent location.
 
-            field_estimate  = obj.position'*obj.QuadModelEst.Q_id*obj.position+obj.QuadModelEst.b_id'*obj.position+obj.QuadModelEst.c_id;
-            field_data_self = field_data.values(1);
-            
-            % We use the following heuristic for identifying the underlying
-            % field which has been observed to work for an inverted gaussian field:
-            % 1. If we have less than three data points, we assume a zero
-            % model since the identfied model wouldn\t be reliable
-            % 2. If we have three to ten data points, we try to identify a
-            % linear model which has three model parameters (grad and bias)
-            % 3. If we have more than ten data points, we identify a
-            % quadratic model which has 6 model parameters (hess, grad and bias)
-            data_size=size(field_data.values,2);
-            if abs(field_estimate-field_data_self) >= obj.field_sensor.noise_bound
-                if (data_size>=3) && (data_size<10)
-                    obj.QuadModelEst=obj.field_sensor.linear_regression(field_data);
-                elseif (data_size>=10)
-                    obj.QuadModelEst=obj.field_sensor.quadratic_regression(field_data);
-                else
-                    obj.QuadModelEst.Q_id  = zeros(obj.dim);
-                    obj.QuadModelEst.b_id  = zeros(obj.dim,1);
-                    obj.QuadModelEst.c_id  = 0;
-                end
-            end
-            grad = 2*obj.QuadModelEst.Q_id*obj.position+obj.QuadModelEst.b_id;
-            hess = 2*obj.QuadModelEst.Q_id;
+%             field_estimate  = obj.position'*obj.QuadModelEst.Q_id*obj.position+obj.QuadModelEst.b_id'*obj.position+obj.QuadModelEst.c_id;
+%             field_data_self = field_data.values(1);
+%             
+%             % We use the following heuristic for identifying the underlying
+%             % field which has been observed to work for an inverted gaussian field:
+%             % 1. If we have less than three data points, we assume a zero
+%             % model since the identfied model wouldn\t be reliable
+%             % 2. If we have three to ten data points, we try to identify a
+%             % linear model which has three model parameters (grad and bias)
+%             % 3. If we have more than ten data points, we identify a
+%             % quadratic model which has 6 model parameters (hess, grad and bias)
+%             data_size=size(field_data.values,2);
+%             if abs(field_estimate-field_data_self) >= obj.field_sensor.noise_bound
+%                 if (data_size>=3) && (data_size<10)
+%                     obj.QuadModelEst=obj.field_sensor.linear_regression(field_data);
+%                 elseif (data_size>=10)
+%                     obj.QuadModelEst=obj.field_sensor.quadratic_regression(field_data);
+%                 else
+%                     obj.QuadModelEst.Q_id  = zeros(obj.dim);
+%                     obj.QuadModelEst.b_id  = zeros(obj.dim,1);
+%                     obj.QuadModelEst.c_id  = 0;
+%                 end
+%             end
+%             grad = 2*obj.QuadModelEst.Q_id*obj.position+obj.QuadModelEst.b_id;
+%             hess = 2*obj.QuadModelEst.Q_id;
 
             % Can get the true gradient and hessians at obj.position
             % for a posterior analysis
             grad=obj.conc_field.get_true_gradient(obj.position);
+            hess=0;
             %hess=obj.conc_field.get_true_hessian(obj.position);
         end