The paper provides a novel approach to control the motion and orientation of a mobile robot using an encoded sequence of arm movements, obtained from the motor imagery indicated by electroencephalographic measurements. The importance of the proposed scheme lies in maintaining secrecy and privacy in control or management of remote robotic systems, as the signals liberated from the user's brain is not accessible to any third party even during the execution phase of the command and hence can find applications in the defense sector. For our demonstration we have successfully differentiated six classes of limb movements from the raw EEG data, encoded the classified signals and used this to control the movements of a Khepera mobile robot. Experiments have been undertaken to study the suitability of the scheme, and the results are promising.